Repository: hive Updated Branches: refs/heads/master 02b2fb5a9 -> 4e3da98d7
HIVE-13841: Orc split generation returns different strategies with cache enabled vs disabled (Prasanth Jayachandran reviewed by Sergey Shelukhin) Project: http://git-wip-us.apache.org/repos/asf/hive/repo Commit: http://git-wip-us.apache.org/repos/asf/hive/commit/4e3da98d Tree: http://git-wip-us.apache.org/repos/asf/hive/tree/4e3da98d Diff: http://git-wip-us.apache.org/repos/asf/hive/diff/4e3da98d Branch: refs/heads/master Commit: 4e3da98d7f05ae29c71bd379c3f59691588c0209 Parents: 02b2fb5 Author: Prasanth Jayachandran <[email protected]> Authored: Fri May 27 16:41:50 2016 -0700 Committer: Prasanth Jayachandran <[email protected]> Committed: Fri May 27 16:41:50 2016 -0700 ---------------------------------------------------------------------- .../hadoop/hive/ql/io/orc/OrcInputFormat.java | 7 ++++--- .../hive/ql/io/orc/TestInputOutputFormat.java | 21 ++++++++++++++++++++ 2 files changed, 25 insertions(+), 3 deletions(-) ---------------------------------------------------------------------- http://git-wip-us.apache.org/repos/asf/hive/blob/4e3da98d/ql/src/java/org/apache/hadoop/hive/ql/io/orc/OrcInputFormat.java ---------------------------------------------------------------------- diff --git a/ql/src/java/org/apache/hadoop/hive/ql/io/orc/OrcInputFormat.java b/ql/src/java/org/apache/hadoop/hive/ql/io/orc/OrcInputFormat.java index 33fe3b6..087207b 100644 --- a/ql/src/java/org/apache/hadoop/hive/ql/io/orc/OrcInputFormat.java +++ b/ql/src/java/org/apache/hadoop/hive/ql/io/orc/OrcInputFormat.java @@ -155,6 +155,7 @@ public class OrcInputFormat implements InputFormat<NullWritable, OrcStruct>, private static final long DEFAULT_MIN_SPLIT_SIZE = 16 * 1024 * 1024; private static final long DEFAULT_MAX_SPLIT_SIZE = 256 * 1024 * 1024; + private static final int DEFAULT_ETL_FILE_THRESHOLD = 100; /** * When picking the hosts for a split that crosses block boundaries, @@ -510,7 +511,7 @@ public class OrcInputFormat implements InputFormat<NullWritable, OrcStruct>, private final int splitStrategyBatchMs; private final long maxSize; private final long minSize; - private final int minSplits; + private final int etlFileThreshold; private final boolean footerInSplits; private final boolean cacheStripeDetails; private final boolean forceThreadpool; @@ -555,7 +556,7 @@ public class OrcInputFormat implements InputFormat<NullWritable, OrcStruct>, cacheStripeDetails = (cacheStripeDetailsSize > 0); - this.minSplits = Math.min(cacheStripeDetailsSize, minSplits); + this.etlFileThreshold = minSplits <= 0 ? DEFAULT_ETL_FILE_THRESHOLD : minSplits; synchronized (Context.class) { if (threadPool == null) { @@ -1938,7 +1939,7 @@ public class OrcInputFormat implements InputFormat<NullWritable, OrcStruct>, deltas, covered, isOriginal, ugi, allowSyntheticFileIds); default: // HYBRID strategy - if (avgFileSize > context.maxSize || totalFiles <= context.minSplits) { + if (avgFileSize > context.maxSize || totalFiles <= context.etlFileThreshold) { return combineOrCreateETLStrategy(combinedCtx, context, fs, dir, baseOrOriginalFiles, deltas, covered, isOriginal, ugi, allowSyntheticFileIds); } else { http://git-wip-us.apache.org/repos/asf/hive/blob/4e3da98d/ql/src/test/org/apache/hadoop/hive/ql/io/orc/TestInputOutputFormat.java ---------------------------------------------------------------------- diff --git a/ql/src/test/org/apache/hadoop/hive/ql/io/orc/TestInputOutputFormat.java b/ql/src/test/org/apache/hadoop/hive/ql/io/orc/TestInputOutputFormat.java index c1ef0e7..52098ae 100644 --- a/ql/src/test/org/apache/hadoop/hive/ql/io/orc/TestInputOutputFormat.java +++ b/ql/src/test/org/apache/hadoop/hive/ql/io/orc/TestInputOutputFormat.java @@ -525,6 +525,27 @@ public class TestInputOutputFormat { } } } + + k = 0; + conf.set("hive.orc.cache.stripe.details.size", "-1"); + for (int c : counts) { + for (int s : sizes) { + final FileSystem fs = generateMockFiles(c, s); + for (int n : numSplits) { + final OrcInputFormat.Context context = new OrcInputFormat.Context( + conf, n); + OrcInputFormat.FileGenerator gen = new OrcInputFormat.FileGenerator( + context, fs, new MockPath(fs, "mock:/a/b"), false, null); + final SplitStrategy splitStrategy = createSplitStrategy(context, gen); + assertTrue( + String.format( + "Split strategy for %d files x %d size for %d splits", c, s, + n), + splitStrategy.getClass().getSimpleName() + .equals(strategyResults[k++])); + } + } + } } @Test
