[SPARK-56010][BUILD] Upgrade snowflake-jdbc to 4.0.2

[dongjoon-hyun]

What changes were proposed in this pull request?

This PR aims to upgrade snowflake-jdbc to 4.0.2.

Why are the changes needed?

To bring the latest bug fixed version for testing:

• https://docs.snowflake.com/en/release-notes/clients-drivers/jdbc-2026#version-4-0-2-mar-12-2026
  • https://github.com/apache/spark/security/dependabot/169

Does this PR introduce any user-facing change?

No. This is a test-only dependency.

How was this patch tested?

Pass the CIs.

Was this patch authored or co-authored using generative AI tooling?

No.

[dongjoon-hyun]

Could you review this Snowflake JDBC Driver CVE patch, @huaxingao ?

$ build/sbt "sql/testOnly *.JDBCSuite"
...
[info] JDBCSuite:
[info] - SELECT * (1 second, 264 milliseconds)
[info] - SELECT * WHERE (simple predicates) (845 milliseconds)
[info] - SELECT COUNT(1) WHERE (predicates) (237 milliseconds)
[info] - SELECT * WHERE (quoted strings) (43 milliseconds)
[info] - SELECT first field (35 milliseconds)
[info] - SELECT first field when fetchsize is two (23 milliseconds)
[info] - SELECT second field (36 milliseconds)
[info] - SELECT second field when fetchsize is two (19 milliseconds)
[info] - SELECT * partitioned (28 milliseconds)
[info] - SELECT WHERE (simple predicates) partitioned (72 milliseconds)
[info] - SELECT second field partitioned (24 milliseconds)
[info] - SPARK-34843: columnPartition should generate the correct stride size and also realign the first partition for better distribution (21 milliseconds)
[info] - overflow of partition bound difference does not give negative stride (23 milliseconds)
[info] - Register JDBC query with renamed fields (10 milliseconds)
[info] - Basic API (30 milliseconds)
[info] - Missing partition columns (47 milliseconds)
[info] - Basic API with FetchSize (101 milliseconds)
[info] - Partitioning via JDBCPartitioningInfo API (22 milliseconds)
[info] - Partitioning via list-of-where-clauses API (18 milliseconds)
[info] - Partitioning on column that might have null values. (84 milliseconds)
[info] - Partitioning on column where numPartitions is zero (40 milliseconds)
[info] - Partitioning on column where numPartitions are more than the number of total rows (42 milliseconds)
[info] - Partitioning on column where lowerBound is equal to upperBound (39 milliseconds)
[info] - Partitioning on column where lowerBound is larger than upperBound (5 milliseconds)
[info] - SELECT * on partitioned table with a nullable partition column (21 milliseconds)
[info] - H2 integral types (37 milliseconds)
[info] - H2 null entries (19 milliseconds)
[info] - H2 string types (42 milliseconds)
[info] - H2 time types (69 milliseconds)
[info] - SPARK-34357: test TIME types (107 milliseconds)
[info] - SPARK-47396: TIME WITHOUT TIME ZONE preferTimestampNTZ (43 milliseconds)
[info] - test DATE types (53 milliseconds)
[info] - test DATE types in cache (63 milliseconds)
[info] - test types for null value (33 milliseconds)
[info] - H2 floating-point types (80 milliseconds)
[info] - SQL query as table name (39 milliseconds)
[info] - Remap types via JdbcDialects (32 milliseconds)
[info] - Map TINYINT to ByteType via JdbcDialects (30 milliseconds)
[info] - Default jdbc dialect registration (1 millisecond)
[info] - quote column names by jdbc dialect (1 millisecond)
[info] - compile filters (66 milliseconds)
[info] - Dialect unregister (1 millisecond)
[info] - Aggregated dialects (1 millisecond)
[info] - Aggregated dialects: isCascadingTruncateTable (0 milliseconds)
[info] - DB2Dialect type mapping (1 millisecond)
[info] - MySQLDialect catalyst type mapping (2 milliseconds)
[info] - SPARK-35446: MySQLDialect type mapping of float (1 millisecond)
[info] - PostgresDialect type mapping (1 millisecond)
[info] - DerbyDialect jdbc type mapping (0 milliseconds)
[info] - SPARK-39604: DerbyDialect catalyst type mapping (1 millisecond)
[info] - OracleDialect jdbc type mapping (0 milliseconds)
[info] - SPARK-42469: OracleDialect Limit query test (1 millisecond)
[info] - MsSqlServerDialect jdbc type mapping (1 millisecond)
[info] - SPARK-28152: MsSqlServerDialect catalyst type mapping (1 millisecond)
[info] - SPARK-42469: MsSqlServerDialect Limit query test (1 millisecond)
[info] - SPARK-42534: DB2Dialect Limit query test !!! IGNORED !!!
[info] - table exists query by jdbc dialect (0 milliseconds)
[info] - truncate table query by jdbc dialect (0 milliseconds)
[info] - SPARK-22880: Truncate table with CASCADE by jdbc dialect (1 millisecond)
[info] - Test DataFrame.where for Date and Timestamp (22 milliseconds)
[info] - SPARK-33867: Test DataFrame.where for LocalDate and Instant (30 milliseconds)
[info] - test credentials in the properties are not in plan output (23 milliseconds)
[info] - test credentials in the connection url are not in the plan output (11 milliseconds)
[info] - hide credentials in create and describe a persistent/temp table (116 milliseconds)
[info] - Hide credentials in show create table (51 milliseconds)
[info] - Replace CatalogUtils.maskCredentials with SQLConf.get.redactOptions (102 milliseconds)
[info] - SPARK 12941: The data type mapping for StringType to Oracle (1 millisecond)
[info] - SPARK-16625: General data types to be mapped to Oracle (0 milliseconds)
[info] - SPARK-15916: JDBC filter operator push down should respect operator precedence (66 milliseconds)
[info] - SPARK-16387: Reserved SQL words are not escaped by JDBC writer (11 milliseconds)
[info] - SPARK-18141: Predicates on quoted column names in the jdbc data source (313 milliseconds)
[info] - SPARK-18419: Fix `asConnectionProperties` to filter case-insensitively (6 milliseconds)
[info] - SPARK-16848: jdbc API throws an exception for user specified schema (3 milliseconds)
[info] - jdbc API support custom schema (37 milliseconds)
[info] - jdbc API custom schema DDL-like strings. (37 milliseconds)
[info] - SPARK-48399: TeradataDialect jdbc data mapping (0 milliseconds)
[info] - SPARK-38846: TeradataDialect catalyst type mapping (0 milliseconds)
[info] - Checking metrics correctness with JDBC (48 milliseconds)
[info] - unsupported types (5 milliseconds)
[info] - SPARK-47394: Convert TIMESTAMP WITH TIME ZONE to TimestampType (136 milliseconds)
[info] - SPARK-19318: Connection properties keys should be case-sensitive. (2 milliseconds)
[info] - SPARK-19318: jdbc data source options should be treated case-insensitive. (71 milliseconds)
[info] - SPARK-23856 Spark jdbc setQueryTimeout option (1 second, 83 milliseconds)
[info] - SPARK-24327 verify and normalize a partition column based on a JDBC resolved schema (19 milliseconds)
[info] - query JDBC option - negative tests (17 milliseconds)
[info] - query JDBC option (69 milliseconds)
[info] - SPARK-22814 support date/timestamp types in partitionColumn (137 milliseconds)
[info] - throws an exception for unsupported partition column types (3 milliseconds)
[info] - SPARK-24288: Enable preventing predicate pushdown (83 milliseconds)
[info] - SPARK-36574: pushDownPredicate=false should prevent push down filters to JDBC data source (91 milliseconds)
[info] - SPARK-26383 throw FAILED_JDBC.CONNECTION if wrong kind of driver to the given url (4 milliseconds)
[info] - support casting patterns for lower/upper bounds of TimestampType (65 milliseconds)
[info] - Add exception when isolationLevel is Illegal (2 milliseconds)
[info] - SPARK-28552: Case-insensitive database URLs in JdbcDialect (0 milliseconds)
[info] - SQLContext.jdbc (deprecated) (92 milliseconds)
[info] - SPARK-32364: JDBCOption constructor (0 milliseconds)
[info] - SPARK-34379: Map JDBC RowID to StringType rather than LongType (802 milliseconds)
[info] - SPARK-39339: Handle TimestampNTZType null values (426 milliseconds)
[info] - SPARK-39339: TimestampNTZType with different local time zones (3 seconds, 160 milliseconds)
[info] - SPARK-41990: Filter with composite name (35 milliseconds)
[info] - SPARK-44866: SnowflakeDialect BOOLEAN type mapping (0 milliseconds)
[info] - SPARK-45139: DatabricksDialect url handling (1 millisecond)
[info] - SPARK-45139: DatabricksDialect catalyst type mapping (0 milliseconds)
[info] - SPARK-45139: DatabricksDialect JDBC type mapping (0 milliseconds)
[info] - SPARK-45425: Mapped TINYINT to ShortType for MsSqlServerDialect (0 milliseconds)
[info] - SPARK-45475: saving a table via JDBC should work with observe API (110 milliseconds)
[info] - SPARK-47496: ServiceLoader support for JDBC dialects (0 milliseconds)
[info] - SPARK-47882: createTableColumnTypes need to be mapped to database types (5 milliseconds)
[info] - SPARK-50666: reading hint test (16 milliseconds)
[info] - FAILED_JDBC.CONNECTION (381 milliseconds)
[info] Run completed in 15 seconds, 303 milliseconds.
[info] Total number of tests run: 111
[info] Suites: completed 1, aborted 0
[info] Tests: succeeded 111, failed 0, canceled 0, ignored 1, pending 0
[info] All tests passed.
[success] Total time: 200 s (0:03:20.0), completed Mar 16, 2026, 1:11:43 PM

[huaxingao]

LGTM

[dongjoon-hyun]

Thank you, @huaxingao ! 😄

[dongjoon-hyun]

Merged to master for Apache Spark 4.2.0.