Description
The current SArg to Filter Conversion ignores the case-sensitivity setting during read and always perform case-sensitive search of fields even when asked to perform case-insensitive match using
options()
.searchArgument(sarg, new String[] color:#6a8759}"RIDX"{color)
.allowSARGToFilter(true)
.isSchemaEvolutionCaseAware(false)
leading to the following exceptions that the field is not found in the schema
java.lang.IllegalArgumentException: Field RIDX not found in struct<f1:bigint,f2:decimal(20,6),f3:bigint,f4:string,ridx:bigint>
at org.apache.orc.impl.ParserUtils.findColumn(ParserUtils.java:399)
at org.apache.orc.impl.ParserUtils.findColumn(ParserUtils.java:337)
at org.apache.orc.impl.ParserUtils.findSubtype(ParserUtils.java:285)
at org.apache.orc.TypeDescription.findSubtype(TypeDescription.java:820)
at org.apache.orc.TypeDescription.findSubtype(TypeDescription.java:814)
at org.apache.orc.impl.filter.leaf.LeafFilterFactory.createLeafVectorFilter(LeafFilterFactory.java:230)
at org.apache.orc.impl.filter.FilterFactory.createSArgFilter(FilterFactory.java:123)
at org.apache.orc.impl.filter.FilterFactory.createBatchFilter(FilterFactory.java:63)
at org.apache.orc.impl.RecordReaderImpl.<init>(RecordReaderImpl.java:284)
SArg to Filter conversion should respect the case-sensitivity option as supplied by the reader.
Attachments
Issue Links
- depends upon
-
ORC-744 LazyIO of non-filter columns
-
- Closed
-
- links to