This is an automated email from the ASF dual-hosted git repository.
cloud-fan pushed a commit to branch branch-4.x
in repository https://gitbox.apache.org/repos/asf/spark.git
The following commit(s) were added to refs/heads/branch-4.x by this push:
new dad313b2802d [SPARK-57295][SQL] Extend whitespace-only path validation
consistency to direct file-path queries
dad313b2802d is described below
commit dad313b2802d900ea9c93baffabb9d9d8c3eb24e
Author: Anurag Kumar Dwivedi <[email protected]>
AuthorDate: Tue Jun 30 21:08:07 2026 +0800
[SPARK-57295][SQL] Extend whitespace-only path validation consistency to
direct file-path queries
## Description
This is a follow-up to PR #56356, which improved validation consistency for
namespace locations by treating whitespace-only values as invalid locations.
### What changes were proposed in this pull request?
This PR extends the same validation behavior to direct file-path queries.
Currently, direct file-path validation checks only for empty strings using
`isEmpty`. Consequently, whitespace-only paths such as `" "`, `"\t"`, and
`"\n"` are not recognized as empty during analysis and may fail later with
datasource-specific errors.
This PR updates the validation to use `SparkStringUtils.isBlank(...)`,
ensuring that whitespace-only paths are treated as invalid and consistently
fail with the standard `INVALID_EMPTY_LOCATION` error.
By doing so, the change aligns direct file-path validation with the
existing namespace location validation logic and improves consistency across
Spark SQL location handling.
### Why are the changes needed?
Currently, validation behavior differs depending on the type of location
being processed:
* Empty paths (`""`) are rejected during analysis with
`INVALID_EMPTY_LOCATION`.
* Whitespace-only paths (`" "`, `"\t"`, `"\n"`) may bypass analysis-time
validation and fail later with datasource-specific errors.
Using `SparkStringUtils.isBlank(...)` ensures consistent handling of all
blank path values across Spark SQL.
### Does this PR introduce any user-facing change?
Yes.
Whitespace-only direct file paths are now rejected during analysis with
`INVALID_EMPTY_LOCATION`, providing behavior consistent with namespace location
validation.
### How was this patch tested?
Added regression test coverage for blank path values, including:
* `""`
* `" "`
* `"\t"`
* `"\n"`
and verified that they consistently fail with `INVALID_EMPTY_LOCATION`.
Jira - https://issues.apache.org/jira/browse/SPARK-57295
Closes #56732 from
AnuragKDwivedi/SPARK-57295-db-location-validation-direct-file-path.
Authored-by: Anurag Kumar Dwivedi <[email protected]>
Signed-off-by: Wenchen Fan <[email protected]>
(cherry picked from commit 41cedca6892132292cf7c2fb6db5747958068e8d)
Signed-off-by: Wenchen Fan <[email protected]>
---
.../org/apache/spark/sql/execution/datasources/rules.scala | 3 ++-
.../test/scala/org/apache/spark/sql/SQLQuerySuite.scala | 14 ++++++++++++++
2 files changed, 16 insertions(+), 1 deletion(-)
diff --git
a/sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/rules.scala
b/sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/rules.scala
index 7122dd52ef1a..3e20900da55b 100644
---
a/sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/rules.scala
+++
b/sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/rules.scala
@@ -44,6 +44,7 @@ import org.apache.spark.sql.types.{ArrayType, DataType,
MapType, StructField, St
import org.apache.spark.sql.util.PartitioningUtils.normalizePartitionSpec
import org.apache.spark.sql.util.SchemaUtils
import org.apache.spark.util.ArrayImplicits._
+import org.apache.spark.util.SparkStringUtils
/**
* Replaces [[UnresolvedRelation]]s if the plan is for direct query on files.
@@ -106,7 +107,7 @@ class ResolveSQLOnFile(sparkSession: SparkSession) extends
Rule[LogicalPlan] {
errorClass = "UNSUPPORTED_DATASOURCE_FOR_DIRECT_QUERY",
messageParameters = Map("dataSourceType" -> ident.head))
}
- if (isFileFormat && ident.last.isEmpty) {
+ if (isFileFormat && SparkStringUtils.isBlank(ident.last)) {
unresolved.failAnalysis(
errorClass = "INVALID_EMPTY_LOCATION",
messageParameters = Map("location" -> ident.last))
diff --git a/sql/core/src/test/scala/org/apache/spark/sql/SQLQuerySuite.scala
b/sql/core/src/test/scala/org/apache/spark/sql/SQLQuerySuite.scala
index da6f6aca2adb..52cc8a3e92af 100644
--- a/sql/core/src/test/scala/org/apache/spark/sql/SQLQuerySuite.scala
+++ b/sql/core/src/test/scala/org/apache/spark/sql/SQLQuerySuite.scala
@@ -1697,6 +1697,20 @@ class SQLQuerySuite extends SharedSparkSession with
AdaptiveSparkPlanHelper
}
assert(e.message.contains("Unsupported data source type for direct query
on files: " +
"org.apache.spark.sql.execution.datasources.jdbc"))
+
+ // Test for empty and whitespace-only paths
+ Seq("", " ", "\t", "\n", "\t\n", " \t ").foreach { file_path =>
+ checkError(
+ exception = intercept[AnalysisException] {
+ sql(s"select id from json.`$file_path`")
+ },
+ condition = "INVALID_EMPTY_LOCATION",
+ parameters = Map("location" -> file_path),
+ queryContext = Array(ExpectedContext(
+ fragment = s"json.`$file_path`",
+ start = 15,
+ stop = 15 + s"json.`$file_path`".length - 1)))
+ }
}
test("SortMergeJoin returns wrong results when using UnsafeRows") {
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
