You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@pig.apache.org by da...@apache.org on 2011/09/14 21:03:13 UTC
svn commit: r1170769 [4/7] - in /pig/branches/branch-0.9: ./
test/e2e/harness/ test/e2e/harness/test/conf/ test/e2e/harness/test/tests/
test/e2e/pig/ test/e2e/pig/conf/ test/e2e/pig/deployers/
test/e2e/pig/drivers/ test/e2e/pig/macros/ test/e2e/pig/tes...
Modified: pig/branches/branch-0.9/test/e2e/pig/tests/nightly.conf
URL: http://svn.apache.org/viewvc/pig/branches/branch-0.9/test/e2e/pig/tests/nightly.conf?rev=1170769&r1=1170768&r2=1170769&view=diff
==============================================================================
--- pig/branches/branch-0.9/test/e2e/pig/tests/nightly.conf (original)
+++ pig/branches/branch-0.9/test/e2e/pig/tests/nightly.conf Wed Sep 14 19:03:12 2011
@@ -29,6 +29,8 @@
$cfg = {
'driver' => 'Pig',
'nummachines' => 5,
+ 'verify_with_pig' => 1,
+ 'verify_pig_version' => 'old',
'groups' => [
{
@@ -38,9 +40,6 @@ $cfg = {
'num' => 1,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
store a into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k;",
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
},
{
'num' => 2,
@@ -54,47 +53,33 @@ g = group f by registration;
h = foreach g generate group, SUM(f.d::contributions);
i = order h by $1;
store i into ':OUTPATH:';\,
- 'sql' => "select registration, sum(contributions)
-from studenttab10k as a full outer join votertab10k as b using (name, age)
-where a.age < 50 and b.age < 50
-group by registration
-order by 2;",
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'sortArgs' => ['-t', ' ', '+1', '-2'],
}
]
},
{
'name' => 'LoaderDefaultDir',
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'tests' => [
{
'num' => 1,
'pig' => q\a = load ':INPATH:/dir/studenttab10k' as (name, age, gpa);
store a into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k;",
},
]
},
{
'name' => 'LoaderPigStorageArg',
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'tests' => [
{
'num' => 1,
'pig' => q\a = load ':INPATH:/singlefile/studentcolon10k' using PigStorage(':') as (name, age, gpa);
store a into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studentcolon10k;",
},
{
# load with control character
'num' => 2,
'pig' => q#a = load ':INPATH:/singlefile/studentctrla10k' using PigStorage('\\u0001') as (name, age, gpa);
store a into ':OUTPATH:';#,
- 'sql' => "select name, age, gpa from studentctrla10k;",
},
{
# load and store with control character
@@ -103,7 +88,7 @@ store a into ':OUTPATH:';#,
store a into ':OUTPATH:.intermediate' using PigStorage('\\u0001');
b = load ':OUTPATH:.intermediate' using PigStorage('\\u0001') as (name, age, gpa);
store b into ':OUTPATH:'; #,
- 'sql' => "select name, age, gpa from studentctrla10k;",
+ 'notmq' => 1,
},
]
},
@@ -120,7 +105,7 @@ b = foreach a generate name, org.apache.
store b into ':OUTPATH:.intermediate' using BinStorage();
c = load ':OUTPATH:.intermediate' using BinStorage();
store c into ':OUTPATH:' using org.apache.pig.test.udf.storefunc.StringStore();\,
- 'sql' => "select name, age, gpa from studenttab10k;",
+ 'notmq' => 1,
},
]
},
@@ -135,59 +120,41 @@ store c into ':OUTPATH:' using org.apach
a = load ':INPATH:/singlefile/textdoc' using TextLoader();
b = foreach a generate TOKENIZE((chararray)$0);
store b into ':OUTPATH:' using org.apache.pig.test.udf.storefunc.StringStore();\,
- 'sql' => "select name from textdoc;",
},
]
},
{
'name' => 'FilterBoolean',
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'tests' => [
{
'num' => 1,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name, age, gpa);
b = filter a by name == 'fred allen' and age > 50;
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where name = 'fred allen' and age > 50;",
},
{
'num' => 2,
'pig' => q\a = load ':INPATH:/dir/studenttab10k' using PigStorage() as (name, age, gpa);
b = filter a by name != 'fred allen' or age < 10;
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where name != 'fred allen' or age < 10;",
},
{
'num' => 3,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name, age, gpa);
b = filter a by not (age == 50);
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where age != 50;",
},
{
'num' => 4,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name, age, gpa);
b = filter a by (age >= 50 or name > 'fred') and (gpa <= 3.0 or name >= 'bob');
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where (age >= 50 or name > 'fred') and (gpa <= 3.0 or name >= 'bob');",
},
{
'num' => 5,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name, age, gpa);
b = filter a by age >= 50 or name > 'fred' and gpa <= 3.0 or name >= 'bob';
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where age >= 50 or name > 'fred' and gpa <= 3.0 or name >= 'bob';",
},
# test filter <= and >= for chararray, int and double
{
@@ -195,9 +162,6 @@ where age >= 50 or name > 'fred' and gpa
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name:chararray, age:int, gpa:double);
b = filter a by age >= 40 and age <=50 and gpa >= 2.0 and gpa <= 3.0 and name >= 'bob' and name <= 'fred';
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where age >= 40 and age <=50 and gpa >= 2.0 and gpa <= 3.0 and name >= 'bob' and name <= 'fred';"
},
# test filter <= and >= for bytearray, long and float
{
@@ -205,9 +169,6 @@ where age >= 40 and age <=50 and gpa >=
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name, age:long, gpa:float);
b = filter a by age >= 40 and age <=50 and gpa >= 2.0f and gpa <= 3.0f and name >= 'bob' and name <= 'fred';
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where age >= 40 and age <=50 and gpa >= 2.0 and gpa <= 3.0 and name >= 'bob' and name <= 'fred';"
},
# test filter < and > for chararray, int and double
{
@@ -215,9 +176,6 @@ where age >= 40 and age <=50 and gpa >=
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name:chararray, age:int, gpa:double);
b = filter a by age > 40 and age <50 and gpa > 2.0 and gpa < 3.0 and name > 'bob' and name < 'fred';
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where age > 40 and age <50 and gpa > 2.0 and gpa < 3.0 and name > 'bob' and name < 'fred';"
},
# test filter < and > for bytearray, long and float
{
@@ -225,9 +183,6 @@ where age > 40 and age <50 and gpa > 2.0
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name, age:long, gpa:float);
b = filter a by age > 40 and age <50 and gpa > 2.0f and gpa < 3.0f and name > 'bob' and name < 'fred';
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where age > 40 and age <50 and gpa > 2.0 and gpa < 3.0 and name > 'bob' and name < 'fred';"
},
# test filter <= and >= for explicit cast for chararray, int and double
{
@@ -235,9 +190,6 @@ where age > 40 and age <50 and gpa > 2.0
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name, age, gpa);
b = filter a by (int)age >= 40 and (int)age <=50 and (double)gpa >= 2.0 and (double)gpa <= 3.0 and (chararray)name >= 'bob' and (chararray)name <= 'fred';
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where age >= 40 and age <=50 and gpa >= 2.0 and gpa <= 3.0 and name >= 'bob' and name <= 'fred';"
},
# test filter <= and >= for explicit cast for bytearray, long and float
{
@@ -245,9 +197,6 @@ where age >= 40 and age <=50 and gpa >=
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name, age, gpa);
b = filter a by (long)age >= 40 and (long)age <=50 and (float)gpa >= 2.0f and (float)gpa <= 3.0f and name >= 'bob' and name <= 'fred';
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where age >= 40 and age <=50 and gpa >= 2.0 and gpa <= 3.0 and name >= 'bob' and name <= 'fred';"
},
# test filter < and > for explicit cast for chararray, int and double
{
@@ -255,9 +204,6 @@ where age >= 40 and age <=50 and gpa >=
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name, age, gpa);
b = filter a by (int)age > 40 and (int)age <50 and (double)gpa > 2.0 and (double)gpa < 3.0 and (chararray)name > 'bob' and (chararray)name < 'fred';
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where age > 40 and age <50 and gpa > 2.0 and gpa < 3.0 and name > 'bob' and name < 'fred';"
},
# test filter < and > for explicit cast for bytearray, long and float
{
@@ -265,9 +211,6 @@ where age > 40 and age <50 and gpa > 2.0
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name, age, gpa);
b = filter a by (long)age > 40 and (long)age <50 and (float)gpa > 2.0f and (float)gpa < 3.0f and name > 'bob' and name < 'fred';
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where age > 40 and age <50 and gpa > 2.0 and gpa < 3.0 and name > 'bob' and name < 'fred';"
},
# test AND with nulls
{
@@ -275,9 +218,6 @@ where age > 40 and age <50 and gpa > 2.0
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' using PigStorage() as (name, age, gpa);
b = filter a by name == 'fred allen' and age > 50;
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studentnulltab10k
-where name = 'fred allen' and age > 50;",
},
# test OR with nulls
{
@@ -285,9 +225,6 @@ where name = 'fred allen' and age > 50;"
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' using PigStorage() as (name, age, gpa);
b = filter a by name != 'fred allen' or age < 10;
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studentnulltab10k
-where name != 'fred allen' or age < 10;",
},
# test with nulls filter <= and >= for chararray, int and double
{
@@ -295,9 +232,6 @@ where name != 'fred allen' or age < 10;"
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' using PigStorage() as (name:chararray, age:int, gpa:double);
b = filter a by age >= 40 and age <=50 and gpa >= 2.0 and gpa <= 3.0 and name >= 'bob' and name <= 'fred';
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studentnulltab10k
-where age >= 40 and age <=50 and gpa >= 2.0 and gpa <= 3.0 and name >= 'bob' and name <= 'fred';"
},
# test with nulls filter < and > for explicit cast for chararray, int and double
{
@@ -305,70 +239,47 @@ where age >= 40 and age <=50 and gpa >=
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' using PigStorage() as (name, age, gpa);
b = filter a by (int)age > 40 and (int)age <50 and (double)gpa > 2.0 and (double)gpa < 3.0 and (chararray)name > 'bob' and (chararray)name < 'fred';
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studentnulltab10k
-where age > 40 and age <50 and gpa > 2.0 and gpa < 3.0 and name > 'bob' and name < 'fred';"
},
],
},
{
'name' => 'FilterEq',
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'tests' => [
{
'num' => 1,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name, age, gpa);
b = filter a by name == 'alice johnson' and age == 64 and gpa == 3.99;
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where name = 'alice johnson' and age = 64 and gpa >= 3.99;", # I know this doesn't match the query above, but for whatever weird reason when you say gpa = 3.99 in postgres it doesn't match.
},
{
'num' => 2,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name, age, gpa);
b = filter a by name > 'fred allen' and age > 40 and gpa > 2.50;
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where name > 'fred allen' and age > 40 and gpa > 2.50;",
},
{
'num' => 3,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name, age, gpa);
b = filter a by name >= 'fred allen' and age >= 40 and gpa >= 2.50;
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where name >= 'fred allen' and age >= 40 and gpa >= 2.50;",
},
{
'num' => 4,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name, age, gpa);
b = filter a by name lt 'fred allen' and age < 40 and gpa < 2.50;
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where name < 'fred allen' and age < 40 and gpa < 2.50;",
},
{
'num' => 5,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name, age, gpa);
b = filter a by name lte 'fred allen' and age <= 40 and gpa <= 2.50;
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where name <= 'fred allen' and age <= 40 and gpa <= 2.50;",
},
{
'num' => 6,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage();
b = filter a by $0 neq 'fred allen' and $1 != '40' and $2 != '2.50';
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where name != 'fred allen' and age != 40 and gpa != 2.50;",
},
# test for filter == for chararray, int and double
{
@@ -376,9 +287,6 @@ where name != 'fred allen' and age != 40
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name:chararray, age:int, gpa:double);
b = filter a by name == 'fred allen' and age == 61 and gpa == 1.42;
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where name = 'fred allen' and age = 61 and cast(gpa as numeric) = 1.42;",
},
# test for filter == for bytearray, long and float
{
@@ -386,9 +294,6 @@ where name = 'fred allen' and age = 61 a
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name, age:long, gpa:float);
b = filter a by name == 'fred allen' and age == 61 and gpa == 1.42f;
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where name = 'fred allen' and age = 61 and cast(gpa as numeric) = 1.42;",
},
# test for filter != for chararray, int and double
{
@@ -396,9 +301,6 @@ where name = 'fred allen' and age = 61 a
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name:chararray, age:int, gpa:double);
b = filter a by $0 != 'fred allen' and $1 != 40 and $2 != 2.50;
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where name != 'fred allen' and age != 40 and gpa != 2.50;",
},
# test for filter != for bytearray, long and float
{
@@ -406,9 +308,6 @@ where name != 'fred allen' and age != 40
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name, age:long, gpa:float);
b = filter a by $0 != 'fred allen' and $1 != 40 and $2 != 2.50f;
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where name != 'fred allen' and age != 40 and gpa != 2.50;",
},
# test for filter == for explicit casts to chararray, int and double
{
@@ -416,9 +315,6 @@ where name != 'fred allen' and age != 40
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name, age, gpa);
b = filter a by (chararray)name == 'fred allen' and (int)age == 61 and (double)gpa == 1.42;
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where name = 'fred allen' and age = 61 and cast(gpa as numeric) = 1.42;",
},
# test for filter == for explicit casts to bytearray, long and float
{
@@ -426,9 +322,6 @@ where name = 'fred allen' and age = 61 a
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name, age, gpa);
b = filter a by name == 'fred allen' and (long)age == 61 and (float)gpa == 1.42f;
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where name = 'fred allen' and age = 61 and cast(gpa as numeric) = 1.42;",
},
# test for filter != for explicit casts to chararray, int and double
{
@@ -436,9 +329,6 @@ where name = 'fred allen' and age = 61 a
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() ;
b = filter a by (chararray)$0 != 'fred allen' and (int)$1 != 40 and (double)$2 != 2.50;
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where name != 'fred allen' and age != 40 and gpa != 2.50;",
},
# test for filter != for explicit casts to bytearray, long and float
{
@@ -446,34 +336,23 @@ where name != 'fred allen' and age != 40
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() ;
b = filter a by $0 != 'fred allen' and (long)$1 != 40 and (float)$2 != 2.50f;
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where name != 'fred allen' and age != 40 and gpa != 2.50;",
},
]
},
{
'name' => 'FilterMatches',
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'tests' => [
{
'num' => 1,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name, age, gpa);
b = filter a by name matches '^fred.*';
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where name like 'fred%';",
},
{
'num' => 2,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage();
b = filter a by not $0 matches '^fred.*';
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, gpa
-from studenttab10k
-where name not like 'fred%';",
},
{
# test for filter on matches for chararray (declared and explicit cast)
@@ -481,54 +360,39 @@ where name not like 'fred%';",
'pig' => q\a = load ':INPATH:/singlefile/votertab10k' using PigStorage() as (name:chararray, age:int, registration, contributions:double);
b = filter a by name matches '^fred.*' and (chararray)registration matches '^dem.*';
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, registration, contributions
-from votertab10k
-where name like 'fred%' and registration like 'dem%';",
},
{
'num' => 4,
'pig' => q\a = load ':INPATH:/singlefile/votertab10k' using PigStorage() as (name:chararray, age:int, registration, contributions:double);
b = filter a by name matches 'f.ed' and (chararray)registration matches 'd.m';
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, registration, contributions
-from votertab10k
-where name like 'f_ed' and registration like 'd_m';",
},
{
'num' => 5,
'pig' => q\a = load ':INPATH:/singlefile/votertab10k' using PigStorage() as (name:chararray, age:int, registration, contributions:double);
b = filter a by name matches 'f[^f]ed.*';
store b into ':OUTPATH:' using PigStorage;\,
- 'sql' => "select name, age, registration, contributions
-from votertab10k
-where name ~ 'f[^f]ed.*';",
},
{
'num' => 6,
'pig' => "a = load ':INPATH:/singlefile/votertab10k' using PigStorage() as (name:chararray, age:int, registration, contributions:double);\nb = filter a by name matches '.*\\\\wan.*';\nstore b into ':OUTPATH:' using PigStorage;",
- 'sql' => "select name, age, registration, contributions from votertab10k where name ~ '.*\\\\wan.*';",
},
{
'num' => 7,
'pig' => "a = load ':INPATH:/singlefile/votertab10k' using PigStorage() as (name:chararray, age:int, registration, contributions:double);\nb = filter a by name matches '^e.*\\\\sc.*';\nstore b into ':OUTPATH:' using PigStorage;",
- 'sql' => "select name, age, registration, contributions from votertab10k where name ~ '^e.*\\\\sc.*';",
},
{
'num' => 8,
'pig' => "a = load ':INPATH:/singlefile/votertab10k' using PigStorage() as (name:chararray, age:int, registration, contributions:double);\nb = filter a by name matches 'ethan white';\nstore b into ':OUTPATH:' using PigStorage;",
- 'sql' => "select name, age, registration, contributions from votertab10k where name ~ 'ethan white';",
},
{
'num' => 9,
'pig' => "a = load ':INPATH:/singlefile/studentnulltab10k' using PigStorage() as (name, age, gpa);\nb = filter a by gpa matches '\\\\d\\\\.45';\nstore b into ':OUTPATH:' using PigStorage;",
- 'sql' => "select name, age, gpa from studentnulltab10k where gpa ~ '\\\\d\\\\.45';",
},
]
},
{
'name' => 'FilterUdf',
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'tests' => [
{
'num' => 1,
@@ -540,8 +404,17 @@ d = filter c by not IsEmpty(a);
e = filter d by not IsEmpty(b);
f = foreach e generate flatten(a), flatten(b);
store f into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, registration, contributions
-from studenttab10k as a join votertab10k as b using (name, age);",
+ },
+ {
+ 'num' => 2,
+ 'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
+ b = load ':INPATH:/singlefile/votertab10k' as (name, age, registration, contributions);
+ c = filter a by age < 50;
+ d = filter b by age < 50;
+ e = cogroup c by (name, age), d by (name, age);
+ f = filter e by COUNT(c)> 0 AND COUNT(d)>0;
+ store f into ':OUTPATH:';\,
+ 'rc' => 0
},
]
},
@@ -555,7 +428,6 @@ from studenttab10k as a join votertab10k
b = group a by name;
c = foreach b generate group, COUNT(a.age);
store c into ':OUTPATH:';\,
- 'sql' => "select name, count(age) from studenttab10k group by name;",
},
{
'num' => 2,
@@ -563,7 +435,6 @@ store c into ':OUTPATH:';\,
b = group a by $0;
c = foreach b generate group, COUNT(a.$1);
store c into ':OUTPATH:';\,
- 'sql' => "select name, count(age) from studenttab10k group by name;",
},
{
'num' => 3,
@@ -571,39 +442,20 @@ store c into ':OUTPATH:';\,
b = group a by (name, age);
c = foreach b generate group.name, group.age, COUNT(a.gpa);
store c into ':OUTPATH:';\,
- 'sql' => "select name, age, count(gpa) from studenttab10k group by name, age;",
},
-# Commented out for 1378834 (COUNT(*) syntax may be wrong, but we should detect that and give an error)
-# {
-# 'num' => 4,
-# 'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
-#b = group a by name;
-#c = foreach b generate group, COUNT(*);
-#store c into ':OUTPATH:';\,
-# 'sql' => "select name, count(*) from studenttab10k group by name;",
-# },
{
'num' => 5,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
b = group a all;
c = foreach b generate COUNT(a.$0);
store c into ':OUTPATH:';\,
- 'sql' => "select count(*) from studenttab10k;",
},
{
- # Results doctored, if you change this query you need to copy the
- # expected results into test/nightly/benchmarks
- # Only until we do sum of ints in ints instead of floats
'num' => 6,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
b = group a by name;
c = foreach b generate group, SUM(a.age);
store c into ':OUTPATH:';\,
- 'sql' => "select name, sum(age)
-from studenttab10k
-group by name;",
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
},
{
'num' => 7,
@@ -611,11 +463,6 @@ group by name;",
b = group a by name;
c = foreach b generate group, SUM(a.gpa);
store c into ':OUTPATH:';\,
- 'sql' => "select name, sum(gpa)
-from studenttab10k
-group by name;",
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
},
{
'num' => 8,
@@ -623,25 +470,13 @@ group by name;",
b = group a by name;
c = foreach b generate group, AVG(a.age);
store c into ':OUTPATH:';\,
- 'sql' => "select name, avg(age)
-from studenttab10k
-group by name;",
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
},
{
- # Results doctored, if you change this query you need to copy the
- # expected results into test/nightly/benchmarks
'num' => 9,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
b = group a by name;
c = foreach b generate group, AVG(a.gpa);
store c into ':OUTPATH:';\,
- 'sql' => "select name, avg(gpa)
-from studenttab10k
-group by name;",
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
},
{
'num' => 10,
@@ -649,11 +484,6 @@ group by name;",
b = group a by name;
c = foreach b generate group, MIN(a.gpa);
store c into ':OUTPATH:';\,
- 'sql' => "select name, min(gpa)
-from studenttab10k
-group by name;",
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
},
{
'num' => 11,
@@ -661,11 +491,6 @@ group by name;",
b = group a by name;
c = foreach b generate group, MAX(a.gpa);
store c into ':OUTPATH:';\,
- 'sql' => "select name, max(gpa)
-from studenttab10k
-group by name;",
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
},
{
'num' => 12,
@@ -673,11 +498,6 @@ group by name;",
b = group a by (name, age);
c = foreach b generate flatten(group), SUM(a.gpa);
store c into ':OUTPATH:';\,
- 'sql' => "select name, age, sum(gpa)
-from studenttab10k
-group by name, age;",
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
},
{
'num' => 13,
@@ -687,11 +507,6 @@ c = load ':INPATH:/singlefile/studenttab
d = cogroup b by group, c by name;
e = foreach d generate flatten(group), SUM(c.gpa), COUNT(c.name);
store e into ':OUTPATH:';\,
- 'sql' => "select name, sum(gpa), count(name)
-from studenttab10k
-group by name;",
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
}
],
},
@@ -704,40 +519,30 @@ group by name;",
b = filter a by name lt 'b';
c = foreach b generate ARITY(name, age, gpa);
store c into ':OUTPATH:';\,
- 'sql' => "select 3 from studenttab10k where name < 'b';",
},
{
- # Results doctored, if you change this query you need to copy the
- # expected results into test/nightly/benchmarks
'num' => 2,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name:chararray, age, gpa);
b = filter a by name lt 'b';
c = foreach b generate TOKENIZE(name);
d = foreach c generate flatten($0);
store d into ':OUTPATH:';\,
- 'sql' => "select name from studenttab10k;",
},
{
- # Results doctored, if you change this query you need to copy the
- # expected results into test/nightly/benchmarks
'num' => 3,
'pig' => q\register :FUNCPATH:/testudf.jar;
a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
b = filter a by name lt 'b';
c = foreach b generate org.apache.pig.test.udf.evalfunc.Swap(name, age);
store c into ':OUTPATH:' using org.apache.pig.test.udf.storefunc.StringStore();\,
- 'sql' => "select age, name from studenttab10k where name < 'b';",
},
{
- # Results doctored, if you change this query you need to copy the
- # expected results into test/nightly/benchmarks
'num' => 4,
'pig' => q\register :FUNCPATH:/testudf.jar;
a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
b = filter a by name lt 'b';
c = foreach b generate org.apache.pig.test.udf.evalfunc.CreateMap((chararray)name, age);
store c into ':OUTPATH:' using org.apache.pig.test.udf.storefunc.StringStore();\,
- 'sql' => "select name, age from studenttab10k where name < 'b';",
}
]
},
@@ -745,8 +550,6 @@ store c into ':OUTPATH:' using org.apach
# TODO User defined grouping function
{
'name' => 'CoGroupFlatten',
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'tests' => [
{
'num' => 1,
@@ -757,9 +560,6 @@ d = filter b by age < 20;
e = cogroup c by name, d by name;
f = foreach e generate flatten (c), flatten(d);
store f into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
-from studenttab10k as a full outer join votertab10k as b using(name)
-where a.age < 20 and b.age < 20;",
},
{
'num' => 2,
@@ -770,9 +570,6 @@ d = filter b by $1 < 20;
e = cogroup c by $0, d by $0;
f = foreach e generate flatten (c), flatten(d);
store f into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
-from studenttab10k as a full outer join votertab10k as b using(name)
-where a.age < 20 and b.age < 20;",
},
{
'num' => 3,
@@ -783,9 +580,6 @@ d = filter b by age < 20;
e = cogroup c by (name, age), d by (name, age);
f = foreach e generate flatten (c), flatten(d);
store f into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
-from studenttab10k as a full outer join votertab10k as b using(name, age)
-where a.age < 20 and b.age < 20;",
},
{
'num' => 4,
@@ -795,9 +589,6 @@ d = filter b by age < 20;
e = cogroup a by (name, age) inner, d by (name, age);
f = foreach e generate flatten (a), flatten(d);
store f into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
-from studenttab10k as a right outer join votertab10k as b using(name, age)
-where a.age < 20;",
},
{
'num' => 5,
@@ -807,9 +598,6 @@ c = filter a by age < 20;
e = cogroup c by (name, age), b by (name, age) inner;
f = foreach e generate flatten (c), flatten(b);
store f into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
-from studenttab10k as a left outer join votertab10k as b using(name, age)
-where b.age < 20;",
},
{
'num' => 6,
@@ -818,8 +606,6 @@ b = load ':INPATH:/singlefile/votertab10
e = cogroup a by (name, age) inner, b by (name, age) inner;
f = foreach e generate flatten (a), flatten(b);
store f into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
-from studenttab10k as a join votertab10k as b using(name, age);",
},
{
# Test cogrouping data loaded from two separate loaders. We don't have any data that can join with studenttab that isn't also loaded with PigStorage, so the
@@ -832,16 +618,13 @@ c = load ':INPATH:/singlefile/votertab10
e = cogroup b by (name, age) inner, c by (name, age) inner;
f = foreach e generate flatten (b), flatten(c);
store f into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
-from studenttab10k as a join votertab10k as b using(name, age);",
+ 'notmq' => 1,
},
]
},
{
'name' => 'CoGroup',
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'tests' => [
{
'num' => 1,
@@ -850,16 +633,11 @@ b = load ':INPATH:/singlefile/votertab10
c = cogroup a by name, b by name;
d = foreach c generate flatten(group), COUNT(a) + COUNT(b);
store d into ':OUTPATH:';\,
- 'sql' => "select a.name, b.name, count(*)
-from studenttab10k as a full outer join votertab10k as b using(name)
-group by a.name, b.name;",
},
]
},
{
'name' => 'Join',
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'tests' => [
{
'num' => 1,
@@ -869,9 +647,6 @@ c = filter a by age < 20;
d = filter b by age < 20;
e = join c by name, d by name;
store e into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
-from studenttab10k as a join votertab10k as b using(name)
-where a.age < 20 and b.age < 20;",
},
{
'num' => 2,
@@ -881,9 +656,6 @@ c = filter a by age < 20;
d = filter b by age < 20;
e = join c by $0, d by $0;
store e into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
-from studenttab10k as a join votertab10k as b using(name)
-where a.age < 20 and b.age < 20;",
},
{
'num' => 3,
@@ -893,9 +665,6 @@ c = filter a by age < 20;
d = filter b by age < 20;
e = join c by (name, age), d by (name, age);
store e into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
-from studenttab10k as a join votertab10k as b using(name, age)
-where a.age < 20 and b.age < 20;",
},
# self join with implict split
# JIRA PIG-429
@@ -905,9 +674,6 @@ where a.age < 20 and b.age < 20;",
b = filter a by $1 > 25;
c = join a by $0, b by $0;
store c into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.gpa
-from studenttab10k as a join studenttab10k as b using(name)
-where b.age > 25;",
},
# join with one input having schema and another without
# JIRA PIG-428
@@ -918,8 +684,6 @@ another = load ':INPATH:/singlefile/stud
c = foreach another generate $0, $1+ 10, $2 + 10.0;
d = join a by $0, c by $0;
store d into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age + 10, b.gpa + 10.0
-from studenttab10k as a join studenttab10k as b using(name);",
},
# self join using fragment replicate join
# no types
@@ -927,10 +691,12 @@ from studenttab10k as a join studenttab1
'num' => 6,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
b = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
-c = join a by name, b by name using "repl";
+c = join a by name, b by name using 'repl';
+store c into ':OUTPATH:';\,
+ 'verify_pig_script' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
+b = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
+c = join a by name, b by name ;
store c into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.gpa
-from studenttab10k as a join studenttab10k as b using(name);",
},
# self join using fragment replicate join
# with types and no cast for join key
@@ -938,10 +704,13 @@ from studenttab10k as a join studenttab1
'num' => 7,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name:chararray, age:int, gpa:double);
b = load ':INPATH:/singlefile/studenttab10k' as (name:chararray, age:int, gpa:double);
-c = join a by name, b by name using "repl";
+c = join a by name, b by name using 'repl';
+store c into ':OUTPATH:';\,
+ 'verify_pig_script' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name:chararray, age:int, gpa:double);
+b = load ':INPATH:/singlefile/studenttab10k' as (name:chararray, age:int, gpa:double);
+c = join a by name, b by name ;
store c into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.gpa
-from studenttab10k as a join studenttab10k as b using(name);",
+
},
# self join using fragment replicate join
# with types and cast for join key
@@ -949,10 +718,13 @@ from studenttab10k as a join studenttab1
'num' => 8,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name:chararray, age:int, gpa:double);
b = load ':INPATH:/singlefile/studenttab10k' as (name:chararray, age:int, gpa);
-c = join a by gpa, b by gpa using "repl";
+c = join a by gpa, b by gpa using 'repl';
store c into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.gpa
-from studenttab10k as a join studenttab10k as b using(gpa);",
+ 'verify_pig_script' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name:chararray, age:int, gpa:double);
+b = load ':INPATH:/singlefile/studenttab10k' as (name:chararray, age:int, gpa);
+c = join a by gpa, b by gpa ;
+store c into ':OUTPATH:';\,
+
},
# left outer join
{
@@ -961,8 +733,6 @@ from studenttab10k as a join studenttab1
b = load ':INPATH:/singlefile/voternulltab10k' as (name:chararray, age:long, registration:chararray, contributions:double);
c = join a by name left outer, b by name;
store c into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
-from studentnulltab10k as a left outer join voternulltab10k as b using(name);",
},
# right outer join
{
@@ -971,8 +741,6 @@ from studentnulltab10k as a left outer j
b = load ':INPATH:/singlefile/voternulltab10k' as (name:chararray, age:long, registration:chararray, contributions:double);
c = join a by name right outer, b by name;
store c into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
-from studentnulltab10k as a right outer join voternulltab10k as b using(name);",
},
# full outer join
{
@@ -981,8 +749,6 @@ from studentnulltab10k as a right outer
b = load ':INPATH:/singlefile/voternulltab10k' as (name:chararray, age:long, registration:chararray, contributions:double);
c = join a by name full outer, b by name;
store c into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
-from studentnulltab10k as a full outer join voternulltab10k as b using(name);",
},
# see PIG-1209 join package now uses internalcachedBag, so every tuple on reduce side in this test will spilled to disk.
{
@@ -994,43 +760,35 @@ from studentnulltab10k as a full outer j
d = filter b by age < 20;
e = join c by name, d by name;
store e into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
- from studenttab10k as a join votertab10k as b using(name) where a.age < 20 and b.age < 20;",
},
]
},
{
'name' => 'Foreach',
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'tests' => [
{
'num' => 1,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
b = foreach a generate *;
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k;",
},
{
'num' => 2,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k';
b = foreach a generate *;
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k;",
},
{
'num' => 3,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
b = foreach a generate name, age;
store b into ':OUTPATH:';\,
- 'sql' => "select name, age from studenttab10k;",
},
{
'num' => 4,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k';
b = foreach a generate $0, $2;
store b into ':OUTPATH:';\,
- 'sql' => "select name, gpa from studenttab10k;",
},
{
# test filter, projection, sort , duplicate elimination
@@ -1046,7 +804,6 @@ d = foreach c {
generate group, flatten(co);
}
store d into ':OUTPATH:';\,
- 'sql' => "select distinct age, gpa from studenttab10k where age < 20 and gpa < 3.0 order by age, gpa;"
},
{
# test flatten for map and scalar
@@ -1055,7 +812,6 @@ store d into ':OUTPATH:';\,
a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
b = foreach a generate flatten(name) as n, flatten(org.apache.pig.test.udf.evalfunc.CreateMap((chararray)name, gpa)) as m;
store b into ':OUTPATH:' using org.apache.pig.test.udf.storefunc.StringStore();\,
- 'sql' => "select name, age, gpa from studenttab10k;",
},
{
# test flatten for UDF that returns bag with multiple tuples with multiple columns
@@ -1064,9 +820,6 @@ store b into ':OUTPATH:' using org.apach
a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
b = foreach a generate name, flatten(org.apache.pig.test.udf.evalfunc.CreateTupleBag(age, gpa)) as foo;
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k union all select name, gpa, age from studenttab10k;",
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
},
{
'num' => 8,
@@ -1074,7 +827,6 @@ store b into ':OUTPATH:';\,
c = group a by name;
d = foreach c generate flatten(group), MAX(a.age) + MIN(a.age);
store d into ':OUTPATH:';\,
- 'sql' => "select name, max(age) + min(age) from studenttab10k group by name;",
},
{
# test filter, projection, sort , duplicate elimination
@@ -1090,7 +842,6 @@ d = foreach c {
generate group, flatten(co);
}
store d into ':OUTPATH:';\,
- 'sql' => "select distinct age, gpa from studenttab10k where age < 20 and gpa >= 3.0 and gpa <= 3.5 order by age, gpa;"
},
{
# test filter, projection, sort , duplicate elimination
@@ -1106,7 +857,6 @@ d = foreach c {
generate group, flatten(co);
}
store d into ':OUTPATH:';\,
- 'sql' => "select distinct age, gpa from studenttab10k where age < 20 and name > 'a' and (gpa = 4.0 or gpa != 2.0) order by age, gpa;"
},
{
# test filter, projection, sort , duplicate elimination
@@ -1119,7 +869,6 @@ c = foreach b {
generate exp1, exp2;
}
store c into ':OUTPATH:';\,
- 'sql' => "select (age + gpa), (age + gpa + age) from studenttab10k where age < 20;"
},
{
# test a udf with no args
@@ -1128,14 +877,11 @@ store c into ':OUTPATH:';\,
a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
b = foreach a generate name, org.apache.pig.test.udf.evalfunc.Fred() as fred;
store b into ':OUTPATH:';\,
- 'sql' => "select name, 'fred' from studenttab10k;"
}
]
},
{
'name' => 'Order',
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'tests' => [
{
'num' => 1,
@@ -1143,7 +889,6 @@ store b into ':OUTPATH:';\,
b = foreach a generate name;
c = order b by name;
store c into ':OUTPATH:';\,
- 'sql' => "select name from studenttab10k order by name;",
'sortArgs' => ['-t', ' ', '+0', '-1'],
},
{
@@ -1152,7 +897,6 @@ store c into ':OUTPATH:';\,
b = foreach a generate $1;
c = order b by $0;
store c into ':OUTPATH:';\,
- 'sql' => "select age from studenttab10k order by age;",
'sortArgs' => ['-t', ' ', '+0', '-1'],
},
{
@@ -1161,7 +905,6 @@ store c into ':OUTPATH:';\,
b = foreach a generate gpa;
c = order b by gpa;
store c into ':OUTPATH:';\,
- 'sql' => "select gpa from studenttab10k order by gpa;",
'sortArgs' => ['-t', ' ', '+0', '-1'],
},
{
@@ -1169,7 +912,6 @@ store c into ':OUTPATH:';\,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k';
b = order a by *;
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k order by 1,2,3;",
'sortArgs' => ['-t', ' '],
},
{
@@ -1178,7 +920,6 @@ store b into ':OUTPATH:';\,
b = foreach a generate name, age;
c = order b by name, age;
store c into ':OUTPATH:';\,
- 'sql' => "select name, age from studenttab10k order by 1,2;",
'sortArgs' => ['-t', ' ', '+0', '-2'],
},
{
@@ -1186,7 +927,6 @@ store c into ':OUTPATH:';\,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k';
c = order a by $0;
store c into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k order by name;",
'sortArgs' => ['-t', ' ', '+0', '-1'],
},
{
@@ -1194,7 +934,6 @@ store c into ':OUTPATH:';\,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k';
c = order a by $1;
store c into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k order by age;",
'sortArgs' => ['-t', ' ', '+1', '-2'],
},
{
@@ -1202,7 +941,6 @@ store c into ':OUTPATH:';\,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k';
c = order a by $0, $1;
store c into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k order by name, age;",
'sortArgs' => ['-t', ' ', '+0', '-2'],
},
{
@@ -1210,7 +948,6 @@ store c into ':OUTPATH:';\,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k';
c = order a by $1, $0;
store c into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k order by age, name;",
'sortArgs' => ['-t', ' ', '+1', '-2', '+0', '-1'],
},
{
@@ -1220,7 +957,6 @@ store c into ':OUTPATH:';\,
a = load ':INPATH:/singlefile/studenttab10k';
c = order a by * using org.apache.pig.test.udf.orderby.OrdDesc;
store c into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k order by name desc, age desc, gpa desc;",
'sortArgs' => ['-t', ' ', '-r'],
},
{
@@ -1230,7 +966,6 @@ store c into ':OUTPATH:';\,
a = load ':INPATH:/singlefile/studenttab10k';
c = order a by $0 using org.apache.pig.test.udf.orderby.OrdDesc;
store c into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k order by name desc;",
'sortArgs' => ['-t', ' ', '-r', '+0', '-1'],
},
{
@@ -1240,7 +975,6 @@ store c into ':OUTPATH:';\,
a = load ':INPATH:/singlefile/studenttab10k';
c = order a by $0, $1 using org.apache.pig.test.udf.orderby.OrdDesc;
store c into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k order by name desc, age desc;",
'sortArgs' => ['-t', ' ', '-r', '+0', '-2'],
},
# ALERT All these tests with inner order bys aren't testing the inner
@@ -1251,7 +985,6 @@ store c into ':OUTPATH:';\,
b = group a by $0;
c = foreach b {c1 = order $1 by $1; generate flatten(c1); };
store c into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k order by name, age;",
},
{
'num' => 14,
@@ -1259,8 +992,6 @@ store c into ':OUTPATH:';\,
b = group a by $0;
c = foreach b {c1 = order $1 by *; generate flatten(c1); };
store c into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k order by name, age, gpa;",
- # 'sortArgs' => ['-t', ' '],
},
{
'num' => 15,
@@ -1269,8 +1000,6 @@ a = load ':INPATH:/singlefile/studenttab
b = group a by $0;
c = foreach b {c1 = order $1 by * using org.apache.pig.test.udf.orderby.OrdDesc; generate flatten(c1); };
store c into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k order by name, age desc, gpa desc;",
- #'sortArgs' => ['-t', ' ', '-r'],
},
{
'num' => 16,
@@ -1279,8 +1008,6 @@ a = load ':INPATH:/singlefile/studenttab
b = group a by $0;
c = foreach b {c1 = order $1 by $1 using org.apache.pig.test.udf.orderby.OrdDesc; generate flatten(c1);};
store c into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k order by name, age desc;",
- #'sortArgs' => ['-t', ' ', '-r', '+0', '-1'],
},
{
'num' => 17,
@@ -1288,9 +1015,6 @@ store c into ':OUTPATH:';\,
b = group a by $0;
c = foreach b {c1 = order $1 by $1; generate flatten(c1), MAX($1.$1); };
store c into ':OUTPATH:';\,
- # This SQL query isn't right, because you can't group by name and have
- # age in the output, like you can in Pig Latin
- 'sql' => "select name, age, gpa, max(age) from studenttab10k group by name, age, gpa order by age;",
},
{
# test to make sure the weighted range patitioning
@@ -1300,15 +1024,12 @@ store c into ':OUTPATH:';\,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k';
b = order a by $1 parallel 100;
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k order by age;",
'sortArgs' => ['-t', ' ', '+1', '-2'],
},
]
},
{
'name' => 'Distinct',
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'tests' => [
{
'num' => 1,
@@ -1316,7 +1037,6 @@ store b into ':OUTPATH:';\,
b = foreach a generate name;
c = distinct b;
store c into ':OUTPATH:';\,
- 'sql' => "select distinct name from studenttab10k ;",
},
{
'num' => 2,
@@ -1324,7 +1044,6 @@ store c into ':OUTPATH:';\,
b = foreach a generate $1;
c = distinct b;
store c into ':OUTPATH:';\,
- 'sql' => "select distinct age from studenttab10k ;",
},
{
'num' => 3,
@@ -1332,14 +1051,12 @@ store c into ':OUTPATH:';\,
b = foreach a generate gpa;
c = distinct b;
store c into ':OUTPATH:';\,
- 'sql' => "select distinct gpa from studenttab10k;",
},
{
'num' => 4,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k';
b = distinct a;
store b into ':OUTPATH:';\,
- 'sql' => "select distinct name, age, gpa from studenttab10k;",
},
{
'num' => 5,
@@ -1347,7 +1064,6 @@ store b into ':OUTPATH:';\,
b = foreach a generate name, age;
c = distinct b;
store c into ':OUTPATH:';\,
- 'sql' => "select distinct name, age from studenttab10k;",
},
{
'num' => 6,
@@ -1355,14 +1071,11 @@ store c into ':OUTPATH:';\,
b = group a by name;
c = foreach b { aa = distinct a.age; generate group, COUNT(aa); }
store c into ':OUTPATH:';\,
- 'sql' => "select name, count(distinct age) from studenttab10k group by name;",
}
]
},
{
'name' => 'Cross',
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'tests' => [
{
'num' => 1,
@@ -1372,9 +1085,6 @@ c = filter a by age < 19 and gpa < 1.0;
d = filter b by age < 19;
e = cross c, d;
store e into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
-from studenttab10k as a cross join votertab10k as b
-where a.age < 19 and b.age < 19 and a.gpa < 1.0;"
},
{
'num' => 2,
@@ -1384,9 +1094,6 @@ c = filter a by age < 19 and gpa < 1.0;
d = filter b by age < 19;
e = cross c, d parallel 10;
store e into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
-from studenttab10k as a cross join votertab10k as b
-where a.age < 19 and b.age < 19 and a.gpa < 1.0;"
},
{
'num' => 3,
@@ -1397,9 +1104,6 @@ c = filter a by age < 19 and gpa < 1.0;
d = filter b by age < 19;
e = cross c, d;
store e into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
-from studenttab10k as a cross join votertab10k as b
-where a.age < 19 and b.age < 19 and a.gpa < 1.0;"
},
{
'num' => 4,
@@ -1411,9 +1115,6 @@ d = filter b by age < 25;
e = cross c, d;
f = filter e by c::age < d::age;
store f into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
-from studenttab10k as a, votertab10k as b
-where a.age < 25 and b.age < 25 and a.age < b.age;"
}
]
},
@@ -1428,9 +1129,6 @@ c = foreach a generate name, age;
d = foreach b generate name, age;
e = union c, d;
store e into ':OUTPATH:';\,
- 'sql' => "select name, age from studenttab10k
-union all
-select name, age from votertab10k;",
},
]
},
@@ -1438,194 +1136,148 @@ select name, age from votertab10k;",
'name' => 'Bincond',
'tests' => [
{
- # Results doctored, if you change this query you need to copy the
- # expected results into test/nightly/benchmarks
- # Only until we do integer arithmetic with integers.
'num' => 1,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
b = foreach a generate name, (name matches 'yuri.*' ? age - 10 : (int)age);
store b into ':OUTPATH:';\,
- 'sql' => "select name, case when name like 'yuri%' then age - 10 else age end from studenttab10k;",
},
]
},
{
'name' => 'Glob',
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'tests' => [
{
'num' => 1,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10?' as (name, age, gpa);
b = filter a by name == 'nick miller';
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k where name = 'nick miller';",
},
{
'num' => 2,
'pig' => q\a = load ':INPATH:/singlefile/st*ttab10k' as (name, age, gpa);
b = filter a by name == 'nick miller';
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k where name = 'nick miller';",
},
{
'num' => 3,
'pig' => q\a = load ':INPATH:/singlefile/studenttab*' as (name, age, gpa);
b = filter a by name == 'nick miller';
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k where name = 'nick miller'
-union all
-select name, age, gpa from studenttab20m where name = 'nick miller';",
},
{
'num' => 4,
'pig' => q\a = load ':INPATH:/singlefile/studenttab???' as (name, age, gpa);
b = filter a by name == 'nick miller';
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k where name = 'nick miller'
-union all
-select name, age, gpa from studenttab20m where name = 'nick miller';",
},
{
'num' => 5,
'pig' => q\a = load ':INPATH:/singlefile/studenttab[1-9]0[km]' as (name, age, gpa);
b = filter a by name == 'nick miller';
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k where name = 'nick miller'
-union all
-select name, age, gpa from studenttab20m where name = 'nick miller';",
},
{
'num' => 6,
'pig' => q\a = load ':INPATH:/singlefile/studenttab[13]0[km]' as (name, age, gpa);
b = filter a by name == 'nick miller';
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k where name = 'nick miller';",
},
{
'num' => 7,
'pig' => q\a = load ':INPATH:/singlefile/studenttab[12]0[a-l]' as (name, age, gpa);
b = filter a by name == 'nick miller';
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k where name = 'nick miller';",
},
{
'num' => 8,
+ 'ignore' => 1, # Need to figure out why this fails
'pig' => q\a = load ':INPATH:/glob/star/*good' as (name, age, gpa);
b = filter a by name == 'nick miller';
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttabsomegood where name = 'nick miller'
-union all
-select name, age, gpa from studenttabmoregood where name = 'nick miller';",
},
{
'num' => 9,
+ 'ignore' => 1, # Need to figure out why this fails
'pig' => q\a = load ':INPATH:/glob/star/*' as (name, age, gpa);
b = filter a by name == 'nick miller';
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttabsomegood where name = 'nick miller'
-union all
-select name, age, gpa from studenttabmoregood where name = 'nick miller'
-union all
-select name, age, gpa from studenttabbad where name = 'nick miller';",
}
]
},
{
'name' => 'Arithmetic',
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'tests' => [
{
'num' => 1,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
c = foreach a generate age + 1, (int)gpa + 1;
store c into ':OUTPATH:';\,
- 'sql' => "select age + 1, floor(gpa) + 1 from studenttab10k;",
},
{
'num' => 2,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
c = foreach a generate (double)age + 1.5, gpa + 1.5;
store c into ':OUTPATH:';\,
- 'sql' => "select age + 1.5, gpa + 1.5 from studenttab10k;",
},
{
'num' => 3,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
c = foreach a generate age - 30, (int)gpa - 3;
store c into ':OUTPATH:';\,
- 'sql' => "select age - 30, floor(gpa) - 3 from studenttab10k;",
},
{
'num' => 4,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
c = foreach a generate (double)age - 30.1, gpa - 3.199;
store c into ':OUTPATH:';\,
- 'sql' => "select age - 30.1, gpa - 3.199 from studenttab10k;",
},
{
'num' => 5,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
c = foreach a generate age * 10, (int)gpa * 2;
store c into ':OUTPATH:';\,
- 'sql' => "select age * 10, floor(gpa) * 2 from studenttab10k;",
},
{
'num' => 6,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
c = foreach a generate (double)age * 10.1, gpa * 2.752342;
store c into ':OUTPATH:';\,
- 'sql' => "select age * 10.1, gpa * 2.752342 from studenttab10k;",
},
{
'num' => 7,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
c = foreach a generate age / 30, (int)gpa / 3;
store c into ':OUTPATH:';\,
- 'sql' => "select floor(cast (age as float) / 30), floor(gpa / 3) from studenttab10k;",
},
{
'num' => 8,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
c = foreach a generate (double)age / 30.323, gpa / 3.22;
store c into ':OUTPATH:';\,
- 'sql' => "select age / 30.323, gpa / 3.22 from studenttab10k;",
},
{
'num' => 9,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
c = foreach a generate 3 * age + gpa / 9.1 - 2;
store c into ':OUTPATH:';\,
- 'sql' => "select 3 * age + gpa / 9.1 - 2 from studenttab10k;",
},
{
'num' => 10,
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' as (name, age, gpa);
c = foreach a generate 3 * (age + gpa) / (9.1 - 2);
store c into ':OUTPATH:';\,
- 'sql' => "select 3 * (age + gpa) / (9.1 - 2) from studenttab10k;",
}
]
},
{
'name' => 'Regression',
'tests' => [
-# {
-# 'num' => 1482458,
-# 'pig' => q\a = load ':INPATH:/singlefile/smallblock_studenttab10k';
-#b = group a all;
-#c = foreach b generate COUNT(a.$0);
-#store c into ':OUTPATH:';\,
-# 'sql' => "select count(*) from studenttab10k;",
-# },
{
'num' => 1459894,
'pig' => q\a = load ':INPATH:/singlefile/reg1459894';
b = group a by $0;
c = foreach b generate group, COUNT(a.$1);
store c into ':OUTPATH:';\,
- 'sql' => "select first, count(*) from reg1459894 group by first;",
},
{
'num' => 97,
@@ -1634,15 +1286,6 @@ b = load ':INPATH:/singlefile/votertab10
e = cogroup a by name, b by name;
f = foreach e generate group, COUNT(a), COUNT(b);
store f into ':OUTPATH:';\,
- 'sql' => "select a.name, a.cnt, b.cnt
-from
- (select name, count(*) as cnt
- from studenttab10k
- group by name) as a,
- (select name, count(*) as cnt
- from votertab10k
- group by name) as b
-where a.name = b.name;",
},
{
'num' => 203,
@@ -1673,9 +1316,6 @@ store c into ':OUTPATH:';
--umped over the lazy brown dog.The quick sly fox j
--mped over the lazy brown dog.The quick sly fox ju
--ped over the lazy brown dog.The quick sly fox jum\,
- 'sql' => "select name, count(*)
-from studenttab10k
-group by name;",
}
]
},
@@ -1686,7 +1326,6 @@ group by name;",
'num' => 1,
'pig' => q\a = load ':INPATH:/singlefile/unicode100';
store a into ':OUTPATH:';\,
- 'sql' => "select name from unicode100;",
},
]
},
@@ -1700,7 +1339,6 @@ store a into ':OUTPATH:';\,
a = load ':INPATH:/singlefile/$fname' using PigStorage() as (name, age, gpa);
b = foreach a generate name;
store b into ':OUTPATH:';\,
- 'sql' => "select name from studenttab10k;"
},
{
# test paramter from command line
@@ -1709,7 +1347,6 @@ store b into ':OUTPATH:';\,
'pig' => q\a = load ':INPATH:/singlefile/$fname' using PigStorage() as (name, age, gpa);
b = foreach a generate name;
store b into ':OUTPATH:';\,
- 'sql' => "select name from studenttab10k;"
},
{
# test paramter from param file
@@ -1718,7 +1355,6 @@ store b into ':OUTPATH:';\,
'pig' => q\a = load ':INPATH:/singlefile/$fname' using PigStorage() as (name, age, gpa);
b = foreach a generate name;
store b into ':OUTPATH:';\,
- 'sql' => "select name from studenttab10k;"
},
{
# test command
@@ -1727,116 +1363,84 @@ store b into ':OUTPATH:';\,
a = load ':INPATH:/singlefile/$cmd' using PigStorage() as (name, age, gpa);
b = foreach a generate name;
store b into ':OUTPATH:';\,
- 'sql' => "select name from studenttab10k;"
},
]
},
{
'name' => 'Types',
- 'sortBenchmark' => 1,
- 'sortResults' => 1,
'tests' => [
{
# constants
'num' => 1,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:int, gpa:double);
b = foreach a generate age + 1 + 0.2f + 253645L, gpa+1;
store b into ':OUTPATH:';\,
- 'sql' => "select age + 1 + cast(0.2 as float) + 253645, gpa + 1 from studentnulltab10k;"
},
{
# NULL and cast
- # Results doctored
'num' => 2,
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:int, gpa:double);
b = foreach a generate (int)((int)gpa/((int)gpa - 1)) as norm_gpa:int;
c = foreach b generate (norm_gpa is null? 0 :norm_gpa);
store c into ':OUTPATH:';\,
- 'sql' => "select case when (cast(gpa as integer) - 1 = 0 or gpa is null) then 0 else cast ((cast(gpa as integer)/(cast(gpa as integer) - 1)) as integer) end from studentnulltab10k;",
# 'expected_err_regex' => "Encountered Warning DIVIDE_BY_ZERO 2387 time.*",
# Driver does currently not support both 'sql' and 'expected_...' verification directives.
},
{
# arithmetic operators and SIZE for int, double and size and concat operators for chararrays
'num' => 3,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:int, gpa:double);
b = foreach a generate age, gpa, age % 25, age + 25, age - 25, age/2, age * 2, SIZE(age), gpa + 10.1, gpa - 1.1 , gpa / 1.2, gpa * 2.5, SIZE(gpa), SIZE(name), CONCAT(name, 'test');
store b into ':OUTPATH:';\,
- 'sql' => "select age, gpa, age % 25, age + 25, age - 25, age/2, age * 2, case when age is null then null else 1 end, gpa + 10.1, gpa - 1.1 , gpa / 1.2, gpa * 2.5, case when gpa is null then null else 1 end, char_length(name), name || 'test' from studentnulltab10k;",
},
{
# arithmetic operators and SIZE for long, float and size and concat operators for bytearrays
'num' => 4,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name, age:long, gpa:float);
b = foreach a generate age, gpa, age % 2L, age + 2500000000L, age - 2500000000L, age/2L, age * 250000000L, SIZE(age), gpa + 10.1f, gpa - 1.1f , gpa / 1.2f, gpa * 2.6f, SIZE(gpa), SIZE(name), CONCAT(name, name);
store b into ':OUTPATH:';\,
- 'sql' => "select age, gpa, age % 2, age + 2500000000, age - 2500000000, age/2, cast((age * (cast(250000000 as bigint))) as bigint), case when age is null then null else 1 end, gpa + 10.1, gpa - 1.1 , gpa / 1.2, gpa * 2.6, case when gpa is null then null else 1 end, char_length(name), name || name from studentnulltab10k;",
},
{
# equlity and implicit cast
'num' => 5,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name, age, gpa);
b = filter a by age == '25' and gpa < 3;
store b into ':OUTPATH:';\,
- 'sql' => "select * from studentnulltab10k where age = 25 and gpa < 3.0;",
},
{
- # sum, min, max, avg for chararray, int and double (declared)
- # WILL NEED DOCTORED RESULTS DUE TO PRECISION DIFFERENCES AND
- # BECAUSE IN PIG CURRENTLY COUNT INCLUDES NULLS - THIS AFFECTS
- # AVG
+ # will need to test against previous version of pig
+ # because in pig currently count includes nulls - this affects
+ # avg
'num' => 6,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:int, gpa:double);
b = group a ALL;
c = foreach b generate SUM(a.age), MIN(a.age), MAX(a.age), AVG(a.age), MIN(a.name), MAX(a.name), SUM(a.gpa), MIN(a.gpa), MAX(a.gpa), AVG(a.gpa);
store c into ':OUTPATH:';\,
- 'sql' => "select sum(age), min(age), max(age), avg(age), min(name), max(name), sum(gpa), min(gpa), max(gpa), avg(gpa) from studentnulltab10k;",
},
{
# sum, min, max, avg for long and float (declared)
- # WILL NEED DOCTORED RESULTS DUE TO PRECISION DIFFERENCES
'num' => 7,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name, age:long, gpa:float);
b = group a ALL;
c = foreach b generate SUM(a.age), MIN(a.age), MAX(a.age), AVG(a.age), SUM(a.gpa), MIN(a.gpa), MAX(a.gpa), AVG(a.gpa);
store c into ':OUTPATH:';\,
- 'sql' => "select sum(age), min(age), max(age), avg(age), sum(gpa), min(gpa), max(gpa), avg(gpa) from studentnulltab10k;",
-
},
{
# Explicit casts - arithmetic operators and SIZE for int, double and size and concat operators for chararrays
'num' => 8,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name, age, gpa);
b = foreach a generate (int)age % 25, (int)age + 25, (int)age - 25, (int)age/2, (int)age * 2, SIZE((int)age), (double)gpa + 10.1, (double)gpa - 1.1 , (double)gpa / 1.2, (double)gpa * 2.5, SIZE((double)gpa), SIZE((chararray)name), CONCAT((chararray)name, 'test');
store b into ':OUTPATH:';\,
- 'sql' => "select age % 25, age + 25, age - 25, age/2, age * 2, case when age is null then null else 1 end, gpa + 10.1, gpa - 1.1 , gpa / 1.2, gpa * 2.5, case when gpa is null then null else 1 end, char_length(name), name || 'test' from studentnulltab10k;",
},
{
# Explicit casts - arithmetic operators and SIZE for long, float
'num' => 9,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name, age, gpa);
b = foreach a generate (long)age, (long)age % 2L, (long)age + 2500000000L, (long)age - 2500000000L, (long)age/2L, (long)age * 250000000L, SIZE((long)age), (float)gpa + 10.1f, (float)gpa - 1.1f , (float)gpa / 1.2f, (float)gpa * 2.6f, SIZE((float)gpa);
store b into ':OUTPATH:';\,
- 'sql' => "select age, age % 2, age + 2500000000, age - 2500000000, age/2, cast((age * (cast(250000000 as bigint))) as bigint), case when age is null then null else 1 end, gpa + 10.1, gpa - 1.1 , gpa / 1.2, gpa * 2.6, case when gpa is null then null else 1 end from studentnulltab10k;",
},
{
# Filter is null for chararray and double and is not null for int
@@ -1846,7 +1450,6 @@ b = filter a by name is null and age is
c = group b ALL;
d = foreach c generate COUNT(b);
store d into ':OUTPATH:';\,
- 'sql' => "select count(*) from studentnulltab10k where name is null and age is not null and gpa is null;",
},
{
# Filter is not null for chararray and double and is null for int
@@ -1856,7 +1459,6 @@ b = filter a by name is not null and age
c = group b ALL;
d = foreach c generate COUNT(b);
store d into ':OUTPATH:';\,
- 'sql' => "select count(*) from studentnulltab10k where name is not null and age is null and gpa is not null;",
},
{
# Filter is null for bytearray and float and is not null for long
@@ -1866,7 +1468,6 @@ b = filter a by name is null and age is
c = group b ALL;
d = foreach c generate COUNT(b);
store d into ':OUTPATH:';\,
- 'sql' => "select count(*) from studentnulltab10k where name is null and age is not null and gpa is null;",
},
{
# Filter is not null for bytearray and float and is null for long
@@ -1876,50 +1477,37 @@ b = filter a by name is not null and age
c = group b ALL;
d = foreach c generate COUNT(b);
store d into ':OUTPATH:';\,
- 'sql' => "select count(*) from studentnulltab10k where name is not null and age is null and gpa is not null;",
},
{
# test that sorting is based on the type for chararray, int and double
'num' => 14,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' =>q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:int, gpa:double);
b = order a by name, age, gpa;
store b into ':OUTPATH:';\,
- 'sql' => "select * from studentnulltab10k order by name, age, gpa;",
'sortArgs' => ['-t', ' ', '+0', '-1', '+1n', '-2'],
},
{
# test that sorting descending is based on the type for chararray, int and double
'num' => 15,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' =>q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:int, gpa:double);
b = order a by name desc, age desc, gpa desc;
store b into ':OUTPATH:';\,
- 'sql' => "select * from studentnulltab10k order by name desc, age desc, gpa desc;",
'sortArgs' => ['-t', ' ', '+0r', '-1', '+1nr', '-2'],
},
{
# test that sorting is based on the type for bytearray, long and float
'num' => 16,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' =>q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name, age:long, gpa:float);
b = order a by name, age, gpa;
store b into ':OUTPATH:';\,
- 'sql' => "select * from studentnulltab10k order by name, age, gpa;",
'sortArgs' => ['-t', ' ', '+0', '-1', '+1n', '-2'],
},
{
# test that sorting descending is based on the type for chararray, age and float
'num' => 17,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' =>q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name, age:long, gpa:float);
b = order a by name desc, age desc, gpa desc;
store b into ':OUTPATH:';\,
- 'sql' => "select * from studentnulltab10k order by name desc, age desc, gpa desc;",
'sortArgs' => ['-t', ' ', '+0r', '-1', '+1nr', '-2'],
},
{
@@ -1928,154 +1516,105 @@ store b into ':OUTPATH:';\,
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:int, gpa:double);
b = foreach a generate 0.123456789123456+0.123456789123456;
store b into ':OUTPATH:';\,
- 'sql' => "select 0.123456789123456+0.123456789123456 from studentnulltab10k;",
},
-# {
- # argument construction for agg functions
-# expressions under agg are currently not supported
-# 'num' => 19,
-# 'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:int, gpa:double);
-#b = group a by name;
-#c = foreach b generate group, SUM(a.age*a.gpa);
-#store c into ':OUTPATH:';\,
-# 'sql' => "select name, sum(age*gpa) from studentnulltab10k group by name;",
-# },
{
# order by string
'num' => 20,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:int, gpa:double);
b = order a by name;
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studentnulltab10k order by name;",
'sortArgs' => ['-t', ' ', '+0', '-1'],
},
{
# order by string desc
'num' => 21,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:int, gpa:double);
b = order a by name desc;
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studentnulltab10k order by name desc;",
'sortArgs' => ['-t', ' ', '+0r', '-1'],
},
{
# order by int
'num' => 22,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:int, gpa:double);
b = order a by age;
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studentnulltab10k order by age;",
'sortArgs' => ['-t', ' ', '+1n', '-2'],
},
{
# order by int desc
'num' => 23,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:int, gpa:double);
b = order a by age desc;
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studentnulltab10k order by age desc;",
'sortArgs' => ['-t', ' ', '+1nr', '-2'],
},
{
# order by long
'num' => 24,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:long, gpa:double);
b = order a by age;
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studentnulltab10k order by age;",
'sortArgs' => ['-t', ' ', '+1n', '-2'],
},
{
# order by long desc
'num' => 25,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:long, gpa:double);
b = order a by age desc;
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studentnulltab10k order by age desc;",
'sortArgs' => ['-t', ' ', '+1nr', '-2'],
},
{
# order by float
'num' => 26,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:int, gpa:float);
b = order a by gpa;
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studentnulltab10k order by gpa;",
'sortArgs' => ['-t', ' ', '-k 3n'],
},
{
# order by float desc
'num' => 27,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:int, gpa:float);
b = order a by gpa desc;
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studentnulltab10k order by gpa desc;",
'sortArgs' => ['-t', ' ', '-k 3nr'],
},
{
# order by double
'num' => 28,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:int, gpa:double);
b = order a by gpa;
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studentnulltab10k order by gpa;",
'sortArgs' => ['-t', ' ', '-k 3n'],
},
{
# order by double desc
'num' => 29,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:int, gpa:double);
b = order a by gpa desc;
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studentnulltab10k order by gpa desc;",
'sortArgs' => ['-t', ' ', '-k 3nr'],
},
{
# order by *
'num' => 30,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:int, gpa:double);
b = order a by *;
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studentnulltab10k order by gpa;",
'sortArgs' => ['-t', ' ', '+0', '-1', '+1n', '-2'],
},
{
# order by * desc
'num' => 31,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:int, gpa:double);
b = order a by * desc;
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studentnulltab10k order by gpa desc;",
'sortArgs' => ['-t', ' ', '+0r', '-1', '+1nr', '-2'],
},
{
'num' => 32,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name:chararray, age:int, gpa:double);
b = load ':INPATH:/singlefile/votertab10k' as (name:chararray, age:int, registration:chararray, contributions:double);
c = filter a by age < 20;
@@ -2083,14 +1622,9 @@ d = filter b by age < 20;
e = cogroup c by name, d by name;
f = foreach e generate flatten (c), flatten(d);
store f into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
-from studenttab10k as a full outer join votertab10k as b using(name)
-where a.age < 20 and b.age < 20;",
},
{
'num' => 33,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name:chararray, age:int, gpa:double);
b = load ':INPATH:/singlefile/votertab10k' as (name:chararray, age:int, registration:chararray, contributions:double);
c = filter a by age < 20;
@@ -2098,14 +1632,9 @@ d = filter b by age < 20;
e = cogroup c by age, d by age;
f = foreach e generate flatten (c), flatten(d);
store f into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
-from studenttab10k as a full outer join votertab10k as b using(age)
-where a.age < 20 and b.age < 20;",
},
{
'num' => 34,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name:chararray, age:long, gpa:double);
b = load ':INPATH:/singlefile/votertab10k' as (name:chararray, age:long, registration:chararray, contributions:double);
c = filter a by age < 20;
@@ -2113,14 +1642,9 @@ d = filter b by age < 20;
e = cogroup c by age, d by age;
f = foreach e generate flatten (c), flatten(d);
store f into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
-from studenttab10k as a full outer join votertab10k as b using(age)
-where a.age < 20 and b.age < 20;",
},
{
'num' => 35,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name:chararray, age:float, gpa:double);
b = load ':INPATH:/singlefile/votertab10k' as (name:chararray, age:float, registration:chararray, contributions:double);
c = filter a by age < 20;
@@ -2128,14 +1652,9 @@ d = filter b by age < 20;
e = cogroup c by age, d by age;
f = foreach e generate flatten (c), flatten(d);
store f into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
-from studenttab10k as a full outer join votertab10k as b using(age)
-where a.age < 20 and b.age < 20;",
},
{
'num' => 36,
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'pig' => q\a = load ':INPATH:/singlefile/studenttab10k' using PigStorage() as (name:chararray, age:double, gpa:double);
b = load ':INPATH:/singlefile/votertab10k' as (name:chararray, age:double, registration:chararray, contributions:double);
c = filter a by age < 20;
@@ -2143,84 +1662,62 @@ d = filter b by age < 20;
e = cogroup c by age, d by age;
f = foreach e generate flatten (c), flatten(d);
store f into ':OUTPATH:';\,
- 'sql' => "select a.name, a.age, a.gpa, b.name, b.age, b.registration, b.contributions
-from studenttab10k as a full outer join votertab10k as b using(age)
-where a.age < 20 and b.age < 20;",
},
{
# NULL and cast
- # Results doctored
'num' => 37,
'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:int, gpa:double);
b = foreach a generate (int)((int)gpa/((int)gpa - 1)) as norm_gpa:int;
c = foreach b generate (norm_gpa is not null? norm_gpa: 0);
store c into ':OUTPATH:';\,
- 'sql' => "select case when (cast(gpa as integer) - 1 = 0 or gpa is null) then 0 else cast ((cast(gpa as integer)/(cast(gpa as integer) - 1)) as integer) end from studentnulltab10k;",
},
-# {
-# Commenting out test case till issue with negative PIG-522 is resolved
+ {
# constants
-# 'num' => 38,
-# 'floatpostprocess' => 1,
-# 'delimiter' => ' ',
-# 'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:int, gpa:double);
-#b = foreach a generate -(age + 1 + 0.2f + 253645L), -(gpa+1);
-#store b into ':OUTPATH:';\,
-# 'sql' => "select -(age + 1 + cast(0.2 as float) + 253645), -(gpa + 1) from studentnulltab10k;"
-# },
+ 'num' => 38,
+ 'pig' => q\a = load ':INPATH:/singlefile/studentnulltab10k' as (name:chararray, age:int, gpa:double);
+b = foreach a generate -(age + 1 + 0.2f + 253645L), -(gpa+1);
+store b into ':OUTPATH:';\,
+ },
]
},
{
'name' => 'Limit',
- 'floatpostprocess' => 1,
'tests' => [
{
- # IMPORTANT NOTE: results need to be stored in benchmarks since
- # we order nulls as < any value
- # while postgreSQL does the opposite
'num' => 1,
- 'delimiter' => ' ',
'pig' =>q\a = load ':INPATH:/singlefile/studentnulltab10k';
b = order a by $0, $1;
c = filter b by $0 > 'a'; -- break the sort/limit optimization
d = limit c 100;
store d into ':OUTPATH:';\,
- 'sql' => "select * from studentnulltab10k where name > 'a' order by name, age limit 100;",
'sortArgs' => ['-t', ' ', '+0', '-1'],
},
{
'num' => 2,
- 'delimiter' => ' ',
'pig' =>q\a = load ':INPATH:/singlefile/studentnulltab10k';
b = order a by $0, $1;
c = limit b 100;
store c into ':OUTPATH:';\,
- 'sql' => "select * from studentnulltab10k order by name, age limit 100;",
'sortArgs' => ['-t', ' ', '+0', '-1'],
},
{
# Make sure that limit higher than number of rows doesn't mess stuff up
'num' => 3,
- 'delimiter' => ' ',
'pig' =>q\a = load ':INPATH:/singlefile/studenttab10k';
b = order a by $0, $1;
c = filter b by $1 < 1000;
d = limit c 100000;
store d into ':OUTPATH:';\,
- 'sql' => "select * from studenttab10k order by name, age;",
},
{
'num' => 4,
- 'delimiter' => ' ',
'pig' =>q\a = load ':INPATH:/singlefile/studentnulltab10k';
b = distinct a;
c = limit b 100;
store c into ':OUTPATH:';\,
- 'sql' => "select distinct * from studentnulltab10k limit 100;",
},
{
'num' => 5,
- 'delimiter' => ' ',
'pig' =>q\a = load ':INPATH:/singlefile/studenttab10k';
b = load ':INPATH:/singlefile/votertab10k';
a1 = foreach a generate $0, $1;
@@ -2228,25 +1725,18 @@ b1 = foreach b generate $0, $1;
c = union a1, b1;
d = limit c 100;
store d into ':OUTPATH:';\,
- 'sql' => "select name, age from studenttab10k
-union all
-select name, age from votertab10k
-limit 100;;",
},
{
'num' => 6,
- 'delimiter' => ' ',
'pig' =>q\A = load ':INPATH:/singlefile/studenttab10k' as (name: chararray, age: int, gpa: float);
B = limit A 40;
C = filter B by age == 40;
D = group C by name;
E = foreach D generate group, COUNT(C);
store E into ':OUTPATH:';\,
- 'sql' => "select name, count(*) from (select * from studenttab10k limit 40) where age = 40 group by name;",
},
{
'num' => 7,
- 'delimiter' => ' ',
'pig' =>q\A = load ':INPATH:/singlefile/studenttab10k' as (name: chararray, age: int, gpa: float);
B = group A by name;
C = foreach B {
@@ -2254,11 +1744,9 @@ C = foreach B {
generate group, COUNT(C1);
}
store C into ':OUTPATH:';\,
- 'sql' => "select name, case when count(*) > 10 then 10 else count(*) end from studenttab10k group by name;",
},
{
'num' => 8,
- 'delimiter' => ' ',
'pig' =>q\A = load ':INPATH:/singlefile/studenttab10k' as (name: chararray, age: int, gpa: float);
B = group A by name;
C = foreach B {
@@ -2268,28 +1756,23 @@ C = foreach B {
}
D = filter C by $1 > 0;
store D into ':OUTPATH:';\,
- 'sql' => "select name, case when count(*) > 10 then 10 else count(*) end from studenttab10k where age < 40 group by name;",
}
]
},
{
'name' => 'Split',
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'tests' => [
{
'num' => 1,
'pig' =>q\a = load ':INPATH:/singlefile/studentnulltab10k';
split a into a1 if $0 > 'm', a2 if $0 <= 'm';
store a1 into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studentnulltab10k where name > 'm';",
},
{
'num' => 2,
'pig' =>q\a = load ':INPATH:/singlefile/studentnulltab10k';
split a into a1 if $0 > 'm', a2 if $0 <= 'm';
store a2 into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studentnulltab10k where name <= 'm';",
},
{
'num' => 3,
@@ -2298,9 +1781,6 @@ split a into a1 if $0 > 'm', a2 if $0 <=
b = cogroup a1 by $1, a2 by $1;
c = foreach b generate flatten(a1), flatten(a2);
store c into ':OUTPATH:';\,
- 'sql' => "select A.name, A.age, A.gpa, B.name, B.age, B.gpa
-from (select * from studenttab10k where name > 'm') as A
- join (select * from studenttab10k where name <= 'm') as B using (age);",
},
{
'num' => 4,
@@ -2309,9 +1789,6 @@ split a into a1 if $0 > 'm', a2 if $0 <=
b = cogroup a1 by $1, a2 by $1;
c = foreach b generate flatten($1), flatten($2);
store c into ':OUTPATH:';\,
- 'sql' => "select A.name, A.age, A.gpa, B.name, B.age, B.gpa
-from (select * from studenttab10k where name > 'm') as A
- join (select * from studenttab10k where name <= 'm') as B using (age);",
},
{
'num' => 5,
@@ -2319,7 +1796,6 @@ from (select * from studenttab10k where
split a into a1 if name > 'm', a2 if name <= 'm';
b = distinct a1;
store b into ':OUTPATH:';\,
- 'sql' => "select distinct name, age, gpa from studentnulltab10k where name > 'm';",
},
{
'num' => 6,
@@ -2327,7 +1803,6 @@ store b into ':OUTPATH:';\,
split a into a1 if age > 50, a2 if age <= 25;
b = order a2 by name;
store b into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studentnulltab10k where age <= 25 order by name;",
'sortArgs' => ['-t', ' ', '+0', '-1'],
},
{
@@ -2336,7 +1811,6 @@ store b into ':OUTPATH:';\,
split a into a1 if name > 'm', a2 if age < 50;
b = distinct a1;
store b into ':OUTPATH:';\,
- 'sql' => "select distinct name, age, gpa from studentnulltab10k where name > 'm';",
},
{
'num' => 8,
@@ -2347,10 +1821,6 @@ b1 = foreach a1 generate name, 2;
c = cogroup b2 by name, b1 by name;
d = foreach c generate flatten(group), COUNT($1), COUNT($2);
store d into ':OUTPATH:';\,
- 'sql' => "select A.name, count(*)
-from (select distinct * from studenttab10k where age > 50) as A
- join (select * from studenttab10k where name < 'm') as B using (name)
-group by A.name;",
},
{
'num' => 9,
@@ -2361,17 +1831,11 @@ b1 = order a1 by name;
c = cogroup b2 by name, b1 by name;
d = foreach c generate flatten(group), COUNT($1), COUNT($2);
store d into ':OUTPATH:';\,
- 'sql' => "select A.name, B.name, count(*)
-from (select distinct * from studenttab10k where age > 50) as A
- join (select * from studenttab10k where name < 'm') as B using (name)
-group by name;",
}
]
},
{
'name' => 'ImplicitSplit',
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'tests' => [
{
'num' => 1,
@@ -2381,9 +1845,6 @@ c = filter a by $2 > 3.0;
d = cogroup b by $0, c by $0;
e = foreach d generate flatten(b), flatten(c);
store e into ':OUTPATH:';\,
- 'sql' => "select A.name, A.age, A.gpa, B.name, B.age, B.gpa
-from (select * from studenttab10k where age > 50) as A
- join (select * from studenttab10k where gpa > 3.0) as B using (name);",
},
{
'num' => 2,
@@ -2394,17 +1855,11 @@ d = cogroup b by name, c by name;
e = foreach d generate flatten(b), flatten(c);
f = filter e by b::age < 75;
store f into ':OUTPATH:';\,
- 'sql' => "select A.name, A.age, A.gpa, B.name, B.age, B.gpa
-from (select * from studenttab10k where age > 50) as A
- join (select * from studenttab10k where gpa > 3.0) as B using (name)
-where A.age < 75;",
}
]
},
{
'name' => 'describe',
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'tests' => [
#JIRA[PIG-373]
{
@@ -2414,14 +1869,11 @@ A = load ':INPATH:/singlefile/studenttab
describe A;
store A into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k;",
},
],
},
{
'name' => 'MissingColumns',
- 'floatpostprocess' => 1,
- 'delimiter' => ' ',
'tests' => [
{
'num' => 1,
@@ -2431,7 +1883,6 @@ B = filter A by age > 50 or extra > 'm';
D = order B by age, extra;
store D into ':OUTPATH:';\,
- 'sql' => "select name, age, gpa from studenttab10k where age > 50;",
'sortArgs' => ['-t', ' ', '+1n', '-2'],
},
{
@@ -2443,7 +1894,6 @@ C = group B by ($0, $2);
D = foreach C generate flatten(group), COUNT($1);
store D into ':OUTPATH:';\,
- 'sql' => "select name, '', count(*) from studenttab10k group by name;",
},
{
'num' => 3,
@@ -2454,7 +1904,6 @@ C = join A by (name, age), B by (name, e
store C into ':OUTPATH:';\,
[... 1851 lines stripped ...]