You are viewing a plain text version of this content. The canonical link for it is here.
Posted to hcatalog-commits@incubator.apache.org by ga...@apache.org on 2012/03/16 22:50:37 UTC
svn commit: r1301812 - in /incubator/hcatalog/branches/branch-0.4:
CHANGES.txt src/test/e2e/hcatalog/conf/default.conf
src/test/e2e/hcatalog/tests/hadoop.conf
src/test/e2e/hcatalog/tests/hive.conf src/test/e2e/hcatalog/tests/pig.conf
Author: gates
Date: Fri Mar 16 22:50:37 2012
New Revision: 1301812
URL: http://svn.apache.org/viewvc?rev=1301812&view=rev
Log:
HCATALOG-306 Need more end-to-end tests
Modified:
incubator/hcatalog/branches/branch-0.4/CHANGES.txt
incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/conf/default.conf
incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/tests/hadoop.conf
incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/tests/hive.conf
incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/tests/pig.conf
Modified: incubator/hcatalog/branches/branch-0.4/CHANGES.txt
URL: http://svn.apache.org/viewvc/incubator/hcatalog/branches/branch-0.4/CHANGES.txt?rev=1301812&r1=1301811&r2=1301812&view=diff
==============================================================================
--- incubator/hcatalog/branches/branch-0.4/CHANGES.txt (original)
+++ incubator/hcatalog/branches/branch-0.4/CHANGES.txt Fri Mar 16 22:50:37 2012
@@ -53,6 +53,8 @@ Release 0.4.0 - Unreleased
HCAT-2 Support nested schema conversion between Hive an Pig (julienledem via hashutosh)
IMPROVEMENTS
+ HCAT-306 Need more end-to-end tests (gates)
+
HCAT-130 Documentation improvements (lefty and gates via gates)
HCAT-266 Upgrade HBase dependency to 0.92 (thw via toffer)
Modified: incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/conf/default.conf
URL: http://svn.apache.org/viewvc/incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/conf/default.conf?rev=1301812&r1=1301811&r2=1301812&view=diff
==============================================================================
--- incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/conf/default.conf (original)
+++ incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/conf/default.conf Fri Mar 16 22:50:37 2012
@@ -62,7 +62,7 @@ $cfg = {
, 'hcat_data_dir' => '/user/hcat/tests/data'
, 'hivehome' => $ENV{'PH_HIVE_HOME'}
, 'hcathome' => $ENV{'HCAT_INSTALL_DIR'}
- , 'hcatalog.jar' => "$ENV{HCAT_JAR},$ENV{HIVE_ROOT}/build/dist/lib/hive-serde-$hive_version.jar,$ENV{HIVE_ROOT}/build/dist/lib/hive-exec-$hive_version.jar,$ENV{PIG_HOME}/pig-0.9.2-withouthadoop.jar,$ENV{HIVE_ROOT}/build/dist/lib/hive-metastore-$hive_version.jar,$ENV{HIVE_ROOT}/build/dist/lib/libfb303-0.7.0.jar,$ENV{HIVE_ROOT}/build/dist/lib/jdo2-api-2.3-ec.jar,$ENV{HCAT_ROOT}/storage-handlers/hbase/build/ivy/lib/hbase-storage-handler/hbase-0.92.0-SNAPSHOT.jar,$ENV{HCAT_ROOT}/storage-handlers/hbase/build/ivy/lib/hbase-storage-handler/zookeeper-3.4.3.jar,$ENV{HCAT_ROOT}/storage-handlers/hbase/build/hbase-storage-handler/hbase-storage-handler-0.1.0.jar,$ENV{HCAT_ROOT}/hive/external/build/dist/lib/hive-hbase-handler-$hive_version.jar"
+ , 'hcatalog.jar' => "$ENV{HCAT_JAR},$ENV{HIVE_ROOT}/build/dist/lib/hive-serde-$hive_version.jar,$ENV{HIVE_ROOT}/build/dist/lib/hive-exec-$hive_version.jar,$ENV{PIG_HOME}/pig-0.9.2-withouthadoop.jar,$ENV{HIVE_ROOT}/build/dist/lib/hive-metastore-$hive_version.jar,$ENV{HIVE_ROOT}/build/dist/lib/libfb303-0.7.0.jar,$ENV{HIVE_ROOT}/build/dist/lib/jdo2-api-2.3-ec.jar,$ENV{HCAT_ROOT}/storage-handlers/hbase/build/ivy/lib/hbase-storage-handler/hbase-0.92.1-SNAPSHOT.jar,$ENV{HCAT_ROOT}/storage-handlers/hbase/build/ivy/lib/hbase-storage-handler/zookeeper-3.4.3.jar,$ENV{HCAT_ROOT}/storage-handlers/hbase/build/hbase-storage-handler/hbase-storage-handler-0.1.0.jar,$ENV{HCAT_ROOT}/hive/external/build/dist/lib/hive-hbase-handler-$hive_version.jar"
#PIG
, 'testconfigpath' => "$ENV{PH_CLUSTER}"
@@ -72,7 +72,7 @@ $cfg = {
, 'pigpath' => "$ENV{PIG_HOME}"
, 'pigjar' => "$ENV{PIG_JAR}" # Pig jar that doesn't have Antlr
, 'oldpigpath' => "$ENV{PH_OLDPIG}"
- , 'additionaljars' => "$ENV{HCAT_ROOT}/build/hcatalog/hcatalog-0.4.0.jar:$ENV{HCAT_ROOT}/hive/external/build/metastore/hive-metastore-$hive_version.jar:$ENV{HCAT_ROOT}/hive/external/build/dist/lib/libthrift.jar:$ENV{HCAT_ROOT}/hive/external/build/dist/lib/hive-exec-$hive_version.jar:$ENV{HCAT_ROOT}/hive/external/build/dist/lib/libfb303.jar:$ENV{HCAT_ROOT}/hive/external/build/dist/lib/jdo2-api-2.3-ec.jar::$ENV{HCAT_ROOT}/storage-handlers/hbase/build/ivy/lib/hbase-storage-handler/hbase-0.92.0-SNAPSHOT.jar:$ENV{HCAT_ROOT}/storage-handlers/hbase/build/ivy/lib/hbase-storage-handler/zookeeper-3.4.3.jar:$ENV{HCAT_ROOT}/storage-handlers/hbase/build/hbase-storage-handler/hbase-storage-handler-0.1.0.jar:$ENV{HCAT_ROOT}/hive/external/build/dist/lib/hive-hbase-handler-$hive_version.jar:$ENV{'HCAT_INSTALL_DIR'}/etc/hcatalog"
+ , 'additionaljars' => "$ENV{HCAT_ROOT}/build/hcatalog/hcatalog-0.4.0.jar:$ENV{HCAT_ROOT}/hive/external/build/metastore/hive-metastore-$hive_version.jar:$ENV{HCAT_ROOT}/hive/external/build/dist/lib/libthrift.jar:$ENV{HCAT_ROOT}/hive/external/build/dist/lib/hive-exec-$hive_version.jar:$ENV{HCAT_ROOT}/hive/external/build/dist/lib/libfb303.jar:$ENV{HCAT_ROOT}/hive/external/build/dist/lib/jdo2-api-2.3-ec.jar::$ENV{HCAT_ROOT}/storage-handlers/hbase/build/ivy/lib/hbase-storage-handler/hbase-0.92.1-SNAPSHOT.jar:$ENV{HCAT_ROOT}/storage-handlers/hbase/build/ivy/lib/hbase-storage-handler/zookeeper-3.4.3.jar:$ENV{HCAT_ROOT}/storage-handlers/hbase/build/hbase-storage-handler/hbase-storage-handler-0.1.0.jar:$ENV{HCAT_ROOT}/hive/external/build/dist/lib/hive-hbase-handler-$hive_version.jar:$ENV{'HCAT_INSTALL_DIR'}/etc/hcatalog"
#HADOOP
, 'hadoopHome' => "$ENV{HCAT_ROOT}/lib"
Modified: incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/tests/hadoop.conf
URL: http://svn.apache.org/viewvc/incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/tests/hadoop.conf?rev=1301812&r1=1301811&r2=1301812&view=diff
==============================================================================
--- incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/tests/hadoop.conf (original)
+++ incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/tests/hadoop.conf Fri Mar 16 22:50:37 2012
@@ -197,7 +197,24 @@ jar :FUNCPATH:/testudf.jar org.apache.hc
,'sql' => q\select name, age, floor(gpa) + 0.1 from all100krc;\
,'floatpostprocess' => 1
,'delimiter' => ' '
- },
+ },{
+ 'num' => 4
+ ,'hcat_prep' => q\
+drop table if exists hadoop_write_4;
+create table hadoop_write_4(
+ name string,
+ age int,
+ gpa double)
+stored as sequencefile;
+\,
+ ,'hadoop' => q\
+jar :FUNCPATH:/testudf.jar org.apache.hcatalog.utils.WriteRC -libjars :HCAT_JAR: :THRIFTSERVER: all100krc hadoop_write_4
+\,
+ ,'result_table' => 'hadoop_write_4'
+ ,'sql' => q\select name, age, floor(gpa) + 0.1 from all100krc;\
+ ,'floatpostprocess' => 1
+ ,'delimiter' => ' '
+ }
],
}, # end g
{
Modified: incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/tests/hive.conf
URL: http://svn.apache.org/viewvc/incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/tests/hive.conf?rev=1301812&r1=1301811&r2=1301812&view=diff
==============================================================================
--- incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/tests/hive.conf (original)
+++ incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/tests/hive.conf Fri Mar 16 22:50:37 2012
@@ -111,6 +111,22 @@ $cfg = {
'name' => 'Hive_Read',
'tests' => [ {
'num' => 1,
+ # float and double columns removed because mysql and hive can't agree
+ # on how to round, even using floor/truncate functions
+ 'sql' => q\select t, si, i, b, s from all100k;\,
+ 'verify_sql' => q\select t, si, i, b, s from all100k;\,
+ 'floatpostprocess' => 1,
+ 'delimiter' => ' ',
+ },{
+ 'num' => 2,
+ # double column removed because mysql and hive can't agree
+ # on how to round, even using floor/truncate functions
+ 'sql' => q\select i, s from all100kjson;\,
+ 'verify_sql' => q\select i, s from all100kjson;\,
+ 'floatpostprocess' => 1,
+ 'delimiter' => ' ',
+ },{
+ 'num' => 3,
'sql' => q\select name, age, floor(gpa) from all100krc;\,
'verify_sql' => q\select name, age, truncate(gpa, 0) from all100krc;\,
'floatpostprocess' => 1,
@@ -121,12 +137,44 @@ $cfg = {
'name' => 'Hive_Write',
'tests' => [ {
'num' => 1,
+ # float and double columns removed because mysql and hive can't agree
+ # on how to round, even using floor/truncate functions
'sql' => q\
drop table if exists hive_write_1;
-create table hive_write_1 (name string, age int, gpa double) stored as rcfile;
-insert into TABLE hive_write_1 select name, age, floor(gpa) + 0.1 from all100krc;\,
+create table hive_write_1 (t tinyint, si smallint, i int, b bigint, s string) row format delimited stored as textfile;
+insert into TABLE hive_write_1 select t, si, i, b, s from all100k;\,
'result_table' => 'hive_write_1',
- 'verify_sql' =>"select name, age, floor(gpa) + 0.1 from all100krc;",
+ 'verify_sql' => q\select t, si, i, b, s from all100k;\,
+ 'floatpostprocess' => 1,
+ 'delimiter' => ' ',
+ },{
+ 'num' => 2,
+ 'sql' => q\
+drop table if exists hive_write_2;
+create table hive_write_2 (name string, age int, gpa double) row format serde 'org.apache.hcatalog.data.JsonSerDe' stored as textfile;
+insert into TABLE hive_write_2 select s, i, 0.1 from all100kjson;\,
+ 'result_table' => 'hive_write_2',
+ 'verify_sql' =>"select s, i, 0.1 from all100kjson;",
+ 'floatpostprocess' => 1,
+ 'delimiter' => ' ',
+ },{
+ 'num' => 3,
+ 'sql' => q\
+drop table if exists hive_write_3;
+create table hive_write_3 (name string, age int, gpa double) stored as rcfile;
+insert into TABLE hive_write_3 select name, age, 1.1 from all100krc;\,
+ 'result_table' => 'hive_write_3',
+ 'verify_sql' =>"select name, age, 1.1 from all100krc;",
+ 'floatpostprocess' => 1,
+ 'delimiter' => ' ',
+ },{
+ 'num' => 4,
+ 'sql' => q\
+drop table if exists hive_write_4;
+create table hive_write_4 (name string, age int, gpa double) stored as sequencefile;
+insert into TABLE hive_write_4 select name, age, 1.1 from studenttab10k;\,
+ 'result_table' => 'hive_write_4',
+ 'verify_sql' =>"select name, age, 1.1 from studenttab10k;",
'floatpostprocess' => 1,
'delimiter' => ' ',
} ]
Modified: incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/tests/pig.conf
URL: http://svn.apache.org/viewvc/incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/tests/pig.conf?rev=1301812&r1=1301811&r2=1301812&view=diff
==============================================================================
--- incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/tests/pig.conf (original)
+++ incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/tests/pig.conf Fri Mar 16 22:50:37 2012
@@ -154,6 +154,34 @@ store b into ':OUTPATH:';\,
,'sql' => q\select name, age from all100krc;\
,'floatpostprocess' => 1
,'delimiter' => ' '
+ },
+ {
+ # A table with one partition in text and one in rc
+ 'num' => 4
+ ,'hcat_prep'=>q?drop table if exists pig_read_4;
+create table pig_read_4 (name string, age int, gpa double) partitioned by (b string) row format delimited fields terminated by '\t' stored as textfile;
+alter table pig_read_4 add partition (b='1') location '/user/hcat/tests/data/studenttab10k';
+alter table pig_read_4 set fileformat rcfile;
+alter table pig_read_4 add partition (b='2') location '/user/hcat/tests/data/all100krc';?
+ ,'pig' => q\a = load 'pig_read_4' using org.apache.hcatalog.pig.HCatLoader();
+b = foreach a generate name, age, b;
+store b into ':OUTPATH:';\,
+ ,'sql' => q\(select name, age, 1 from studenttab10k)
+ union all
+ (select name, age, 2 from all100krc);\
+ },
+ {
+ # Read from a table in the non-default database
+ 'num' => 5
+ ,'hcat_prep'=>q?create database if not exists pig_db_1;
+drop table if exists pig_db_1.pig_read_5;
+create table pig_db_1.pig_read_5 (name string, age int, gpa double) partitioned by (b string) row format delimited fields terminated by '\t' stored as textfile;
+use pig_db_1;
+alter table pig_read_5 add partition (b='1') location '/user/hcat/tests/data/studenttab10k';?
+ ,'pig' => q\a = load 'pig_db_1.pig_read_5' using org.apache.hcatalog.pig.HCatLoader();
+b = foreach a generate name, age, b;
+store b into ':OUTPATH:';\,
+ ,'sql' => q\select name, age, 1 from studenttab10k;\
}
],
}, # end g
@@ -208,10 +236,82 @@ store b into 'pig_write_3' using org.apa
,'result_table' => 'pig_write_3'
,'floatpostprocess' => 1
,'delimiter' => ' '
+ },
+ {
+ # Store in a sequence file
+ 'num' => 4
+ ,'hcat_prep'=>q\drop table if exists pig_write_4;
+create table pig_write_4(
+ name string,
+ age int,
+ gpa double)
+stored as sequencefile;
+\
+ ,'pig' => q\a = load 'studenttab10k' using org.apache.hcatalog.pig.HCatLoader();
+b = foreach a generate name, age, 0.1;
+c = foreach b generate name, age, $2 as gpa;
+store c into 'pig_write_4' using org.apache.hcatalog.pig.HCatStorer();\,
+ ,'sql' => q\select name, age, 0.1 from studenttab10k;\
+ ,'result_table' => 'pig_write_4'
+ ,'floatpostprocess' => 1
+ ,'delimiter' => ' '
+ },
+ {
+ # Write to a table in the non-default database
+ 'num' => 5
+ ,'hcat_prep'=>q?create database if not exists pig_db_1;
+create table if not exists pig_db_1.pig_write_5 (name string, age int) row format delimited fields terminated by '\t' stored as textfile;?
+ ,'pig' => q\a = load 'studenttab10k' using org.apache.hcatalog.pig.HCatLoader();
+b = foreach a generate name, age;
+store b into 'pig_db_1.pig_write_5' using org.apache.hcatalog.pig.HCatStorer();\,
+ ,'sql' => q\select name, age from studenttab10k;\
+ ,'result_table' => 'pig_db_1.pig_write_5'
}
],
}, # end g
{
+ 'name' => 'Pig_Change_Schema',
+ 'tests' => [
+ {
+ # I don't like this, I'm using one test to setup for the next. But I don't know how else to do this.
+ 'num' => 1
+ ,'hcat_prep'=>q\drop table if exists pig_change_schema_1;
+create table pig_change_schema_1 (name string) partitioned by (ds string) STORED AS TEXTFILE;\
+ ,'pig' => q\a = load 'studentparttab30k' using org.apache.hcatalog.pig.HCatLoader();
+b = filter a by ds == '20110924';
+c = foreach b generate name;
+store c into 'pig_change_schema_1' using org.apache.hcatalog.pig.HCatStorer('ds=20110924');\,
+ ,'result_table' => 'pig_change_schema_1'
+ ,'sql' => q\select name, ds from studentparttab30k where ds='20110924';\
+ },
+ {
+ # I don't like this, I'm using one test to setup for the next. But I don't know how else to do this.
+ 'num' => 2
+ ,'depends_on' => 'Pig_Change_Schema_1'
+ ,'hcat_prep'=>q\alter table pig_change_schema_1 add columns (age int);\
+ ,'pig' => q\a = load 'studentparttab30k' using org.apache.hcatalog.pig.HCatLoader();
+b = filter a by ds == '20110925';
+c = foreach b generate name, age;
+store c into 'pig_change_schema_1' using org.apache.hcatalog.pig.HCatStorer('ds=20110925');\,
+ ,'result_table' => 'pig_change_schema_1'
+ ,'sql' => q\(select name, '', ds from studentparttab30k where ds='20110924')
+ union all
+ (select name, age, ds from studentparttab30k where ds = '20110925');\
+ },
+ {
+ # I don't like this, I'm using one test to setup for the next. But I don't know how else to do this.
+ 'num' => 3
+ , 'depends_on' => 'Pig_Change_Schema_2'
+ ,'pig' => q\a = load 'pig_change_schema_1' using org.apache.hcatalog.pig.HCatLoader();
+c = foreach a generate name, age, ds;
+store c into ':OUTPATH:';\
+ ,'sql' => q\(select name, '', ds from studentparttab30k where ds='20110924')
+ union all
+ (select name, age, ds from studentparttab30k where ds = '20110925');\
+ }
+ ],
+ },
+ {
'name' => 'Pig_HBase',
'tests' => [
{