Github user HyukjinKwon commented on a diff in the pull request:
https://github.com/apache/spark/pull/17527#discussion_r110317272
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/HadoopFsRelation.scala
---
@@ -52,7 +54,11 @@ case class HadoopFsRelation(
val schema: StructType = {
val getColName: (StructField => String) =
- if (sparkSession.sessionState.conf.caseSensitiveAnalysis) _.name
else _.name.toLowerCase
+ if (sparkSession.sessionState.conf.caseSensitiveAnalysis) {
+ _.name
+ } else {
+ _.name.toLowerCase(Locale.ROOT)
+ }
--- End diff --
I think we should leave this out. It seems `dataSchema` means the schema
from the source which is exposed to users. I think this could cause a problem.
For example as below:
Before
```scala
import collection.mutable
import org.apache.spark.sql.types._
java.util.Locale.setDefault(new java.util.Locale("tr"))
val partitionSchema: StructType = StructType(StructField("I", StringType)
:: Nil)
val dataSchema: StructType = StructType(StructField("ı", StringType) ::
Nil)
val getColName: (StructField => String) = _.name.toLowerCase
val overlappedPartCols = mutable.Map.empty[String, StructField]
partitionSchema.foreach { partitionField =>
if (dataSchema.exists(getColName(_) == getColName(partitionField))) {
overlappedPartCols += getColName(partitionField) -> partitionField
}
}
val schema = StructType(dataSchema.map(f =>
overlappedPartCols.getOrElse(getColName(f), f)) ++
partitionSchema.filterNot(f =>
overlappedPartCols.contains(getColName(f))))
schema.fieldNames
```
prints
```scala
Array[String] = Array(I)
```
After
```scala
import collection.mutable
import org.apache.spark.sql.types._
java.util.Locale.setDefault(new java.util.Locale("tr"))
val partitionSchema: StructType = StructType(StructField("I", StringType)
:: Nil)
val dataSchema: StructType = StructType(StructField("ı", StringType) ::
Nil)
val getColName: (StructField => String) =
_.name.toLowerCase(java.util.Locale.ROOT)
val overlappedPartCols = mutable.Map.empty[String, StructField]
partitionSchema.foreach { partitionField =>
if (dataSchema.exists(getColName(_) == getColName(partitionField))) {
overlappedPartCols += getColName(partitionField) -> partitionField
}
}
val schema = StructType(dataSchema.map(f =>
overlappedPartCols.getOrElse(getColName(f), f)) ++
partitionSchema.filterNot(f =>
overlappedPartCols.contains(getColName(f))))
schema.fieldNames
```
prints
```scala
Array[String] = Array(ı, I)
```
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
