You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@impala.apache.org by dh...@apache.org on 2017/05/19 16:22:04 UTC
[1/2] incubator-impala git commit: IMPALA-5137: Support TIMESTAMPs in
Kudu range predicate DDL
Repository: incubator-impala
Updated Branches:
refs/heads/master 24c77f194 -> d6e612f5c
IMPALA-5137: Support TIMESTAMPs in Kudu range predicate DDL
Adds support in DDL for timestamps in Kudu range partition syntax.
For convenience, strings can be specified with or without
explicit casts to TIMESTAMP.
E.g.
create table ts_ranges (ts timestamp primary key, i int)
partition by range (
partition '2009-01-02 00:00:00' <= VALUES < '2009-01-03 00:00:00'
) stored as kudu
Range bounds are converted to Kudu UNIXTIME_MICROS during
analysis.
Testing: Adds FE and EE tests.
Change-Id: Iae409b6106c073b038940f0413ed9d5859daaeff
Reviewed-on: http://gerrit.cloudera.org:8080/6849
Reviewed-by: Matthew Jacobs <mj...@cloudera.com>
Tested-by: Impala Public Jenkins
Project: http://git-wip-us.apache.org/repos/asf/incubator-impala/repo
Commit: http://git-wip-us.apache.org/repos/asf/incubator-impala/commit/6226e597
Tree: http://git-wip-us.apache.org/repos/asf/incubator-impala/tree/6226e597
Diff: http://git-wip-us.apache.org/repos/asf/incubator-impala/diff/6226e597
Branch: refs/heads/master
Commit: 6226e597020b6950d7f2f2d43bf54d944e194bda
Parents: 24c77f1
Author: Matthew Jacobs <mj...@cloudera.com>
Authored: Wed May 10 16:28:58 2017 -0700
Committer: Impala Public Jenkins <im...@gerrit.cloudera.org>
Committed: Fri May 19 00:41:46 2017 +0000
----------------------------------------------------------------------
.../apache/impala/analysis/RangePartition.java | 48 +++++++++++--
.../java/org/apache/impala/util/KuduUtil.java | 8 ++-
.../apache/impala/analysis/AnalyzeDDLTest.java | 26 +++++++
.../queries/QueryTest/kudu_alter.test | 42 ++++++++++++
.../queries/QueryTest/kudu_create.test | 72 ++++++++++++++++++++
5 files changed, 188 insertions(+), 8 deletions(-)
----------------------------------------------------------------------
http://git-wip-us.apache.org/repos/asf/incubator-impala/blob/6226e597/fe/src/main/java/org/apache/impala/analysis/RangePartition.java
----------------------------------------------------------------------
diff --git a/fe/src/main/java/org/apache/impala/analysis/RangePartition.java b/fe/src/main/java/org/apache/impala/analysis/RangePartition.java
index b441a78..5ff00c7 100644
--- a/fe/src/main/java/org/apache/impala/analysis/RangePartition.java
+++ b/fe/src/main/java/org/apache/impala/analysis/RangePartition.java
@@ -17,18 +17,20 @@
package org.apache.impala.analysis;
+import java.math.BigInteger;
import java.util.List;
+
+import org.apache.impala.catalog.Type;
+import org.apache.impala.common.AnalysisException;
+import org.apache.impala.common.InternalException;
import org.apache.impala.common.Pair;
+import org.apache.impala.thrift.TRangePartition;
+import org.apache.impala.util.KuduUtil;
import com.google.common.base.Joiner;
import com.google.common.base.Preconditions;
-import com.google.common.collect.ImmutableList;
import com.google.common.collect.Lists;
-import org.apache.impala.common.AnalysisException;
-import org.apache.impala.util.KuduUtil;
-import org.apache.impala.thrift.TRangePartition;
-
/**
* Represents a range partition of a Kudu table.
*
@@ -50,12 +52,18 @@ import org.apache.impala.thrift.TRangePartition;
* are inclusive (true) or exclusive (false).
*/
public class RangePartition implements ParseNode {
+
+ // Upper and lower bound exprs contain literals of the target column type post-analysis.
+ // For TIMESTAMPs those are Kudu UNIXTIME_MICROS, i.e. int64s.
private final List<Expr> lowerBound_;
- private final boolean lowerBoundInclusive_;
private final List<Expr> upperBound_;
+ private final boolean lowerBoundInclusive_;
private final boolean upperBoundInclusive_;
private final boolean isSingletonRange_;
+ // Set true when this partition has been analyzed.
+ private boolean isAnalyzed_ = false;
+
private RangePartition(List<Expr> lowerBoundValues, boolean lowerBoundInclusive,
List<Expr> upperBoundValues, boolean upperBoundInclusive) {
Preconditions.checkNotNull(lowerBoundValues);
@@ -112,10 +120,14 @@ public class RangePartition implements ParseNode {
public void analyze(Analyzer analyzer, List<ColumnDef> partColDefs)
throws AnalysisException {
+ // Reanalyzing not supported because TIMESTAMPs are converted to BIGINT (unixtime
+ // micros) in place.
+ Preconditions.checkArgument(!isAnalyzed_);
analyzeBoundaryValues(lowerBound_, partColDefs, analyzer);
if (!isSingletonRange_) {
analyzeBoundaryValues(upperBound_, partColDefs, analyzer);
}
+ isAnalyzed_ = true;
}
private void analyzeBoundaryValues(List<Expr> boundaryValues,
@@ -158,6 +170,20 @@ public class RangePartition implements ParseNode {
}
org.apache.impala.catalog.Type colType = pkColumn.getType();
Preconditions.checkState(KuduUtil.isSupportedKeyType(colType));
+
+ // Special case string literals in timestamp columns for convenience.
+ if (literal.getType().isStringType() && colType.isTimestamp()) {
+ // Add an explicit cast to TIMESTAMP
+ Expr e = new CastExpr(new TypeDef(Type.TIMESTAMP), literal);
+ e.analyze(analyzer);
+ literal = LiteralExpr.create(e, analyzer.getQueryCtx());
+ Preconditions.checkNotNull(literal);
+ if (literal.isNullLiteral()) {
+ throw new AnalysisException(String.format("Range partition value %s cannot be " +
+ "cast to target TIMESTAMP partitioning column.", value.toSql()));
+ }
+ }
+
org.apache.impala.catalog.Type literalType = literal.getType();
if (!org.apache.impala.catalog.Type.isImplicitlyCastable(literalType, colType,
true)) {
@@ -171,6 +197,16 @@ public class RangePartition implements ParseNode {
literal = LiteralExpr.create(castLiteral, analyzer.getQueryCtx());
}
Preconditions.checkNotNull(literal);
+
+ if (colType.isTimestamp()) {
+ try {
+ long unixTimeMicros = KuduUtil.timestampToUnixTimeMicros(analyzer, literal);
+ literal = new NumericLiteral(BigInteger.valueOf(unixTimeMicros), Type.BIGINT);
+ } catch (InternalException e) {
+ throw new AnalysisException(
+ "Error converting timestamp in range definition: " + toSql(), e);
+ }
+ }
return literal;
}
http://git-wip-us.apache.org/repos/asf/incubator-impala/blob/6226e597/fe/src/main/java/org/apache/impala/util/KuduUtil.java
----------------------------------------------------------------------
diff --git a/fe/src/main/java/org/apache/impala/util/KuduUtil.java b/fe/src/main/java/org/apache/impala/util/KuduUtil.java
index 9fc13b7..8b61c88 100644
--- a/fe/src/main/java/org/apache/impala/util/KuduUtil.java
+++ b/fe/src/main/java/org/apache/impala/util/KuduUtil.java
@@ -154,6 +154,10 @@ public class KuduUtil {
checkCorrectType(literal.isSetString_literal(), type, colName, literal);
key.addString(pos, literal.getString_literal().getValue());
break;
+ case UNIXTIME_MICROS:
+ checkCorrectType(literal.isSetInt_literal(), type, colName, literal);
+ key.addLong(pos, literal.getInt_literal().getValue());
+ break;
default:
throw new ImpalaRuntimeException("Key columns not supported for type: "
+ type.toString());
@@ -196,7 +200,7 @@ public class KuduUtil {
}
}
- public static Long timestampToUnixTimeMicros(Analyzer analyzer, Expr timestampExpr)
+ public static long timestampToUnixTimeMicros(Analyzer analyzer, Expr timestampExpr)
throws AnalysisException, InternalException {
Preconditions.checkArgument(timestampExpr.isAnalyzed());
Preconditions.checkArgument(timestampExpr.isConstant());
@@ -330,7 +334,7 @@ public class KuduUtil {
}
public static boolean isSupportedKeyType(org.apache.impala.catalog.Type type) {
- return type.isIntegerType() || type.isStringType();
+ return type.isIntegerType() || type.isStringType() || type.isTimestamp();
}
/**
http://git-wip-us.apache.org/repos/asf/incubator-impala/blob/6226e597/fe/src/test/java/org/apache/impala/analysis/AnalyzeDDLTest.java
----------------------------------------------------------------------
diff --git a/fe/src/test/java/org/apache/impala/analysis/AnalyzeDDLTest.java b/fe/src/test/java/org/apache/impala/analysis/AnalyzeDDLTest.java
index 48b15c7..0858774 100644
--- a/fe/src/test/java/org/apache/impala/analysis/AnalyzeDDLTest.java
+++ b/fe/src/test/java/org/apache/impala/analysis/AnalyzeDDLTest.java
@@ -2366,6 +2366,32 @@ public class AnalyzeDDLTest extends FrontendTestBase {
AnalysisError("create table tab (i int, x int primary key) partition by hash(x) " +
"partitions 8 sort by(i) stored as kudu", "SORT BY is not supported for Kudu " +
"tables.");
+
+ // Range partitions with TIMESTAMP
+ AnalyzesOk("create table ts_ranges (ts timestamp primary key) " +
+ "partition by range (partition cast('2009-01-01 00:00:00' as timestamp) " +
+ "<= VALUES < '2009-01-02 00:00:00') stored as kudu");
+ AnalyzesOk("create table ts_ranges (ts timestamp primary key) " +
+ "partition by range (partition value = cast('2009-01-01 00:00:00' as timestamp" +
+ ")) stored as kudu");
+ AnalyzesOk("create table ts_ranges (ts timestamp primary key) " +
+ "partition by range (partition value = '2009-01-01 00:00:00') " +
+ "stored as kudu");
+ AnalyzesOk("create table ts_ranges (id int, ts timestamp, primary key(id, ts))" +
+ "partition by range (partition value = (9, cast('2009-01-01 00:00:00' as " +
+ "timestamp))) stored as kudu");
+ AnalyzesOk("create table ts_ranges (id int, ts timestamp, primary key(id, ts))" +
+ "partition by range (partition value = (9, '2009-01-01 00:00:00')) " +
+ "stored as kudu");
+ AnalysisError("create table ts_ranges (ts timestamp primary key, i int)" +
+ "partition by range (partition '2009-01-01 00:00:00' <= VALUES < " +
+ "'NOT A TIMESTAMP') stored as kudu",
+ "Range partition value 'NOT A TIMESTAMP' cannot be cast to target TIMESTAMP " +
+ "partitioning column.");
+ AnalysisError("create table ts_ranges (ts timestamp primary key, i int)" +
+ "partition by range (partition 100 <= VALUES < 200) stored as kudu",
+ "Range partition value 100 (type: TINYINT) is not type " +
+ "compatible with partitioning column 'ts' (type: TIMESTAMP).");
}
@Test
http://git-wip-us.apache.org/repos/asf/incubator-impala/blob/6226e597/testdata/workloads/functional-query/queries/QueryTest/kudu_alter.test
----------------------------------------------------------------------
diff --git a/testdata/workloads/functional-query/queries/QueryTest/kudu_alter.test b/testdata/workloads/functional-query/queries/QueryTest/kudu_alter.test
index 369eecf..20c2c56 100644
--- a/testdata/workloads/functional-query/queries/QueryTest/kudu_alter.test
+++ b/testdata/workloads/functional-query/queries/QueryTest/kudu_alter.test
@@ -413,3 +413,45 @@ select count(*) from external_tbl
---- TYPES
BIGINT
====
+---- QUERY
+create table ts_ranges (ts timestamp primary key, i int)
+partition by range (
+ partition cast('2009-01-01 00:00:00' as timestamp) <= VALUES <
+ cast('2009-01-02 00:00:00' as timestamp)
+) stored as kudu
+---- RESULTS
+====
+---- QUERY
+show range partitions ts_ranges
+---- RESULTS
+'2009-01-01T00:00:00.000000Z <= VALUES < 2009-01-02T00:00:00.000000Z'
+---- TYPES
+STRING
+====
+---- QUERY
+alter table ts_ranges add range partition
+cast('2009-01-02 00:00:00' as timestamp) <= VALUES <
+cast('2009-01-03 00:00:00' as timestamp)
+---- RESULTS
+====
+---- QUERY
+show range partitions ts_ranges
+---- RESULTS
+'2009-01-01T00:00:00.000000Z <= VALUES < 2009-01-02T00:00:00.000000Z'
+'2009-01-02T00:00:00.000000Z <= VALUES < 2009-01-03T00:00:00.000000Z'
+---- TYPES
+STRING
+====
+---- QUERY
+alter table ts_ranges drop range partition
+cast('2009-01-02 00:00:00' as timestamp) <= VALUES <
+cast('2009-01-03 00:00:00' as timestamp)
+---- RESULTS
+====
+---- QUERY
+show range partitions ts_ranges
+---- RESULTS
+'2009-01-01T00:00:00.000000Z <= VALUES < 2009-01-02T00:00:00.000000Z'
+---- TYPES
+STRING
+====
http://git-wip-us.apache.org/repos/asf/incubator-impala/blob/6226e597/testdata/workloads/functional-query/queries/QueryTest/kudu_create.test
----------------------------------------------------------------------
diff --git a/testdata/workloads/functional-query/queries/QueryTest/kudu_create.test b/testdata/workloads/functional-query/queries/QueryTest/kudu_create.test
index b7db12f..704c868 100644
--- a/testdata/workloads/functional-query/queries/QueryTest/kudu_create.test
+++ b/testdata/workloads/functional-query/queries/QueryTest/kudu_create.test
@@ -144,3 +144,75 @@ X, I1, I2, I3, I4, VALS, VALF, VALD, VALB
---- TYPES
INT,TINYINT,SMALLINT,INT,BIGINT,STRING,FLOAT,DOUBLE,BOOLEAN
====
+---- QUERY
+# Overlapping ranges are rejected by the Kudu client
+create table ts_ranges (ts timestamp primary key, i int)
+partition by range (
+ partition cast('2009-01-02 00:00:00' as timestamp) <= VALUES <
+ cast('2009-01-03 00:00:00' as timestamp),
+ partition cast('2009-01-02 10:00:00' as timestamp) <= VALUES
+) stored as kudu
+---- CATCH
+NonRecoverableException: overlapping range partitions: first range partition: 2009-01-02T00:00:00.000000Z <= VALUES < 2009-01-03T00:00:00.000000Z, second range partition: VALUES >= 2009-01-02T10:00:00.000000Z
+====
+---- QUERY
+# Creates a range partitioned Kudu table with a timestamp PK. Note that nanoseconds
+# are rounded (same behavior as when writing timestamp values to Kudu).
+create table ts_ranges (ts timestamp primary key, i int)
+partition by range (
+ partition nanoseconds_add(cast('2009-01-01 00:00:00' as timestamp), 999) <= VALUES <
+ nanoseconds_add(cast('2009-01-02 00:00:00' as timestamp), 1),
+ partition cast('2009-01-02 00:00:00' as timestamp) <= VALUES <
+ cast('2009-01-03 00:00:00' as timestamp),
+ partition '2009-01-03 00:00:00' <= VALUES
+) stored as kudu
+---- RESULTS
+====
+---- QUERY
+show range partitions ts_ranges
+---- RESULTS
+'2009-01-01T00:00:00.000001Z <= VALUES < 2009-01-02T00:00:00.000000Z'
+'2009-01-02T00:00:00.000000Z <= VALUES < 2009-01-03T00:00:00.000000Z'
+'VALUES >= 2009-01-03T00:00:00.000000Z'
+---- TYPES
+STRING
+====
+---- QUERY
+create table ts_ranges_ctas
+primary key (ts)
+partition by range (
+ partition VALUES < '2009-01-02 00:00:00',
+ partition '2009-01-02 00:00:00' <= VALUES <
+ '2009-01-03 00:00:00',
+ partition '2009-01-03 00:00:00' < VALUES,
+ partition VALUE = ('2009-01-03 00:00:00')
+)
+stored as kudu
+as select timestamp_col ts, id from functional.alltypestiny;
+---- RESULTS
+'Inserted 8 row(s)'
+====
+---- QUERY
+show range partitions ts_ranges_ctas
+---- RESULTS
+'VALUES < 2009-01-02T00:00:00.000000Z'
+'2009-01-02T00:00:00.000000Z <= VALUES < 2009-01-03T00:00:00.000000Z'
+'VALUE = 2009-01-03T00:00:00.000000Z'
+'VALUES >= 2009-01-03T00:00:00.000001Z'
+---- TYPES
+STRING
+====
+---- QUERY
+select * from ts_ranges_ctas order by id
+---- RESULTS
+2009-01-01 00:00:00,0
+2009-01-01 00:01:00,1
+2009-02-01 00:00:00,2
+2009-02-01 00:01:00,3
+2009-03-01 00:00:00,4
+2009-03-01 00:01:00,5
+2009-04-01 00:00:00,6
+2009-04-01 00:01:00,7
+---- TYPES
+TIMESTAMP,INT
+====
[2/2] incubator-impala git commit: IMPALA-5180: Don't use
non-deterministic exprs in partition pruning
Posted by dh...@apache.org.
IMPALA-5180: Don't use non-deterministic exprs in partition pruning
Non-deterministic exprs which evaluate as constant should not be
used during HDFS partition pruning. We consider Exprs which have no
SlotRefs as bound by default, and thus we end up trying to apply
them indisrciminately. Constant propagation makes this situation
easier to run into and the behavior is rather unexpected.
The fix for now is to explicitly disallow non-deterministic Exprs
in partition pruning.
Change-Id: I91054c6bf017401242259a1eff5e859085285546
Reviewed-on: http://gerrit.cloudera.org:8080/6575
Reviewed-by: Alex Behm <al...@cloudera.com>
Tested-by: Impala Public Jenkins
Project: http://git-wip-us.apache.org/repos/asf/incubator-impala/repo
Commit: http://git-wip-us.apache.org/repos/asf/incubator-impala/commit/d6e612f5
Tree: http://git-wip-us.apache.org/repos/asf/incubator-impala/tree/d6e612f5
Diff: http://git-wip-us.apache.org/repos/asf/incubator-impala/diff/d6e612f5
Branch: refs/heads/master
Commit: d6e612f5c7c800bce98dc35b2e04dd893d1b6f86
Parents: 6226e59
Author: Zach Amsden <za...@cloudera.com>
Authored: Thu Apr 6 23:41:11 2017 +0000
Committer: Impala Public Jenkins <im...@gerrit.cloudera.org>
Committed: Fri May 19 08:50:25 2017 +0000
----------------------------------------------------------------------
.../apache/impala/analysis/PartitionSet.java | 2 +-
.../impala/planner/HdfsPartitionPruner.java | 3 +-
.../org/apache/impala/planner/HdfsScanNode.java | 1 +
.../queries/PlannerTest/hdfs.test | 42 ++++++++++++++++----
4 files changed, 39 insertions(+), 9 deletions(-)
----------------------------------------------------------------------
http://git-wip-us.apache.org/repos/asf/incubator-impala/blob/d6e612f5/fe/src/main/java/org/apache/impala/analysis/PartitionSet.java
----------------------------------------------------------------------
diff --git a/fe/src/main/java/org/apache/impala/analysis/PartitionSet.java b/fe/src/main/java/org/apache/impala/analysis/PartitionSet.java
index d5f0e70..3ff4ed7 100644
--- a/fe/src/main/java/org/apache/impala/analysis/PartitionSet.java
+++ b/fe/src/main/java/org/apache/impala/analysis/PartitionSet.java
@@ -75,7 +75,7 @@ public class PartitionSet extends PartitionSpecBase {
}
// Make sure every conjunct only contains partition slot refs.
- if(!e.isBoundBySlotIds(partitionSlots)) {
+ if (!e.isBoundBySlotIds(partitionSlots)) {
throw new AnalysisException("Partition exprs cannot contain non-partition " +
"column(s): " + e.toSql() + ".");
}
http://git-wip-us.apache.org/repos/asf/incubator-impala/blob/d6e612f5/fe/src/main/java/org/apache/impala/planner/HdfsPartitionPruner.java
----------------------------------------------------------------------
diff --git a/fe/src/main/java/org/apache/impala/planner/HdfsPartitionPruner.java b/fe/src/main/java/org/apache/impala/planner/HdfsPartitionPruner.java
index cc27528..95ef68c 100644
--- a/fe/src/main/java/org/apache/impala/planner/HdfsPartitionPruner.java
+++ b/fe/src/main/java/org/apache/impala/planner/HdfsPartitionPruner.java
@@ -108,7 +108,8 @@ public class HdfsPartitionPruner {
Iterator<Expr> it = conjuncts.iterator();
while (it.hasNext()) {
Expr conjunct = it.next();
- if (conjunct.isBoundBySlotIds(partitionSlots_)) {
+ if (conjunct.isBoundBySlotIds(partitionSlots_) &&
+ !conjunct.contains(Expr.IS_NONDETERMINISTIC_BUILTIN_FN_PREDICATE)) {
// Check if the conjunct can be evaluated from the partition metadata.
// Use a cloned conjunct to rewrite BetweenPredicates and allow
// canEvalUsingPartitionMd() to fold constant expressions without modifying
http://git-wip-us.apache.org/repos/asf/incubator-impala/blob/d6e612f5/fe/src/main/java/org/apache/impala/planner/HdfsScanNode.java
----------------------------------------------------------------------
diff --git a/fe/src/main/java/org/apache/impala/planner/HdfsScanNode.java b/fe/src/main/java/org/apache/impala/planner/HdfsScanNode.java
index 296dfb4..a3286b4 100644
--- a/fe/src/main/java/org/apache/impala/planner/HdfsScanNode.java
+++ b/fe/src/main/java/org/apache/impala/planner/HdfsScanNode.java
@@ -485,6 +485,7 @@ public class HdfsScanNode extends ScanNode {
List<SlotId> slotIds = Lists.newArrayList();
conjunct.getIds(tupleIds, slotIds);
+ if (slotIds.size() == 0) continue;
Preconditions.checkState(tupleIds.size() == 1);
if (slotIds.size() != 1) continue;
http://git-wip-us.apache.org/repos/asf/incubator-impala/blob/d6e612f5/testdata/workloads/functional-planner/queries/PlannerTest/hdfs.test
----------------------------------------------------------------------
diff --git a/testdata/workloads/functional-planner/queries/PlannerTest/hdfs.test b/testdata/workloads/functional-planner/queries/PlannerTest/hdfs.test
index 5ed6f6b..9b9a678 100644
--- a/testdata/workloads/functional-planner/queries/PlannerTest/hdfs.test
+++ b/testdata/workloads/functional-planner/queries/PlannerTest/hdfs.test
@@ -1016,14 +1016,42 @@ PLAN-ROOT SINK
00:SCAN HDFS [functional.alltypes]
partitions=0/24 files=0 size=0B
====
-# IMPALA-4592: Test that we bail on evaluating non-deterministic predicates when trying
-# to prune partitions.
-select * from functional.alltypes where rand() > 100
+# IMPALA-4592: Test random predicates that reference a partition column.
+select * from functional.alltypes where rand() > year
---- PLAN
-not implemented: Unsupported non-deterministic predicate: rand() > 100
+PLAN-ROOT SINK
+|
+00:SCAN HDFS [functional.alltypes]
+ partitions=24/24 files=24 size=478.45KB
+ predicates: year < rand()
====
-# IMPALA-4592: Same as above but the predicate references a partition column.
-select * from functional.alltypes where rand() > year
+# IMPALA-5180: Test that predicates not touching a partition column are ignored in
+# partition pruning
+select count(*) from
+ (select * from functional.alltypes) T
+where random() = 100
+---- PLAN
+PLAN-ROOT SINK
+|
+01:AGGREGATE [FINALIZE]
+| output: count(*)
+|
+00:SCAN HDFS [functional.alltypes]
+ partitions=24/24 files=24 size=478.45KB
+ predicates: random() = 100
+====
+# Same as above, with a column that gets constant folded away
+select count(*) from
+ (select * from functional.alltypes) T
+where (coalesce(NULL, T.int_col) + random() * T.int_col = 100 OR
+ coalesce(NULL, T.int_col) + T.int_col = 20) and T.int_col = 1
---- PLAN
-not implemented: Unsupported non-deterministic predicate: year < rand()
+PLAN-ROOT SINK
+|
+01:AGGREGATE [FINALIZE]
+| output: count(*)
+|
+00:SCAN HDFS [functional.alltypes]
+ partitions=24/24 files=24 size=478.45KB
+ predicates: functional.alltypes.int_col = 1, FALSE OR 1 + random() * 1 = 100
====