diff options
author | Cheng Lian <lian@databricks.com> | 2015-08-12 20:01:34 +0800 |
---|---|---|
committer | Cheng Lian <lian@databricks.com> | 2015-08-12 20:01:34 +0800 |
commit | 3ecb3794302dc12d0989f8d725483b2cc37762cf (patch) | |
tree | b8c3a132482fe273a71f3f9bb2235bebc395744e /sql/core/src/test/README.md | |
parent | 9d0822455ddc8d765440d58c463367a4d67ef456 (diff) | |
download | spark-3ecb3794302dc12d0989f8d725483b2cc37762cf.tar.gz spark-3ecb3794302dc12d0989f8d725483b2cc37762cf.tar.bz2 spark-3ecb3794302dc12d0989f8d725483b2cc37762cf.zip |
[SPARK-9407] [SQL] Relaxes Parquet ValidTypeMap to allow ENUM predicates to be pushed down
This PR adds a hacky workaround for PARQUET-201, and should be removed once we upgrade to parquet-mr 1.8.1 or higher versions.
In Parquet, not all types of columns can be used for filter push-down optimization. The set of valid column types is controlled by `ValidTypeMap`. Unfortunately, in parquet-mr 1.7.0 and prior versions, this limitation is too strict, and doesn't allow `BINARY (ENUM)` columns to be pushed down. On the other hand, `BINARY (ENUM)` is commonly seen in Parquet files written by libraries like `parquet-avro`.
This restriction is problematic for Spark SQL, because Spark SQL doesn't have a type that maps to Parquet `BINARY (ENUM)` directly, and always converts `BINARY (ENUM)` to Catalyst `StringType`. Thus, a predicate involving a `BINARY (ENUM)` is recognized as one involving a string field instead and can be pushed down by the query optimizer. Such predicates are actually perfectly legal except that it fails the `ValidTypeMap` check.
The workaround added here is relaxing `ValidTypeMap` to include `BINARY (ENUM)`. I also took the chance to simplify `ParquetCompatibilityTest` a little bit when adding regression test.
Author: Cheng Lian <lian@databricks.com>
Closes #8107 from liancheng/spark-9407/parquet-enum-filter-push-down.
Diffstat (limited to 'sql/core/src/test/README.md')
-rw-r--r-- | sql/core/src/test/README.md | 16 |
1 files changed, 6 insertions, 10 deletions
diff --git a/sql/core/src/test/README.md b/sql/core/src/test/README.md index 3dd9861b48..421c2ea4f7 100644 --- a/sql/core/src/test/README.md +++ b/sql/core/src/test/README.md @@ -6,23 +6,19 @@ The following directories and files are used for Parquet compatibility tests: . ├── README.md # This file ├── avro -│ ├── parquet-compat.avdl # Testing Avro IDL -│ └── parquet-compat.avpr # !! NO TOUCH !! Protocol file generated from parquet-compat.avdl +│ ├── *.avdl # Testing Avro IDL(s) +│ └── *.avpr # !! NO TOUCH !! Protocol files generated from Avro IDL(s) ├── gen-java # !! NO TOUCH !! Generated Java code ├── scripts -│ └── gen-code.sh # Script used to generate Java code for Thrift and Avro +│ ├── gen-avro.sh # Script used to generate Java code for Avro +│ └── gen-thrift.sh # Script used to generate Java code for Thrift └── thrift - └── parquet-compat.thrift # Testing Thrift schema + └── *.thrift # Testing Thrift schema(s) ``` -Generated Java code are used in the following test suites: - -- `org.apache.spark.sql.parquet.ParquetAvroCompatibilitySuite` -- `org.apache.spark.sql.parquet.ParquetThriftCompatibilitySuite` - To avoid code generation during build time, Java code generated from testing Thrift schema and Avro IDL are also checked in. -When updating the testing Thrift schema and Avro IDL, please run `gen-code.sh` to update all the generated Java code. +When updating the testing Thrift schema and Avro IDL, please run `gen-avro.sh` and `gen-thrift.sh` accordingly to update generated Java code. ## Prerequisites |