fix: replace() returns wrong result for empty-string search

[wattt3]

Which issue does this PR close?

Closes #3344.

Rationale for this change

What changes are included in this PR?

• Added a SparkStringReplace UDF that handles the empty search string case by inserting the replacement between every character and at boundaries.
• Updated QueryPlanSerde to use the new spark_replace instead of DataFusion's replace.

How are these changes tested?

• Rust unit test for empty search string behavior including null handling.
• Spark integration test in CometStringExpressionSuite comparing results with Spark's output.

[andygrove]

Could you add these tests in the new sql file testing approach instead?

https://datafusion.apache.org/comet/contributor-guide/sql-file-tests.html

[andygrove]

Thanks @wattt3. LGTM. I left a comment about moving the tests to the new sql file based approach. I will be happy to approve the PR once that is done.

[andygrove]

There is a test failure:

2026-02-04T16:17:16.3295718Z - replace with empty search string *** FAILED *** (166 milliseconds)
2026-02-04T16:17:16.3296584Z   Results do not match for query:
2026-02-04T16:17:16.3299606Z   Timezone: sun.util.calendar.ZoneInfo[id="America/Los_Angeles",offset=-28800000,dstSavings=3600000,useDaylight=true,transitions=311,lastRule=java.util.SimpleTimeZone[id=America/Los_Angeles,offset=-28800000,dstSavings=3600000,useDaylight=true,startYear=0,startMode=3,startMonth=2,startDay=8,startDayOfWeek=1,startTime=7200000,startTimeMode=0,endMode=3,endMonth=10,endDay=1,endDayOfWeek=1,endTime=7200000,endTimeMode=0]]
2026-02-04T16:17:16.3302580Z   Timezone Env: 
2026-02-04T16:17:16.3302956Z   
2026-02-04T16:17:16.3303402Z   == Parsed Logical Plan ==
2026-02-04T16:17:16.3304730Z   Project [replace(col#89982, , x) AS replace(col, , x)#89993, replace(col#89982, , ) AS replace(col, , )#89994]
2026-02-04T16:17:16.3306566Z   +- SubqueryAlias spark_catalog.default.test
2026-02-04T16:17:16.3307192Z      +- Relation spark_catalog.default.test[col#89982] parquet
2026-02-04T16:17:16.3307674Z   
2026-02-04T16:17:16.3307942Z   == Analyzed Logical Plan ==
2026-02-04T16:17:16.3308425Z   replace(col, , x): string, replace(col, , ): string
2026-02-04T16:17:16.3309283Z   Project [replace(col#89982, , x) AS replace(col, , x)#89993, replace(col#89982, , ) AS replace(col, , )#89994]
2026-02-04T16:17:16.3310070Z   +- SubqueryAlias spark_catalog.default.test
2026-02-04T16:17:16.3310839Z      +- Relation spark_catalog.default.test[col#89982] parquet
2026-02-04T16:17:16.3311330Z   
2026-02-04T16:17:16.3311631Z   == Optimized Logical Plan ==
2026-02-04T16:17:16.3312688Z   Project [replace(col#89982, , x) AS replace(col, , x)#89993, replace(col#89982, , ) AS replace(col, , )#89994]
2026-02-04T16:17:16.3313560Z   +- Relation spark_catalog.default.test[col#89982] parquet
2026-02-04T16:17:16.3314032Z   
2026-02-04T16:17:16.3314277Z   == Physical Plan ==
2026-02-04T16:17:16.3314578Z   *(1) CometColumnarToRow
2026-02-04T16:17:16.3315628Z   +- CometProject [replace(col, , x)#89993, replace(col, , )#89994], [replace(col#89982, , x) AS replace(col, , x)#89993, replace(col#89982, , ) AS replace(col, , )#89994]
2026-02-04T16:17:16.3318449Z      +- CometScan [native_iceberg_compat] parquet spark_catalog.default.test[col#89982] Batched: true, DataFilters: [], Format: CometParquet, Location: InMemoryFileIndex(1 paths)[file:/__w/datafusion-comet/datafusion-comet/spark/spark-warehouse/test], PartitionFilters: [], PushedFilters: [], ReadSchema: struct<col:string>
2026-02-04T16:17:16.3320372Z   
2026-02-04T16:17:16.3320776Z   == Results ==
2026-02-04T16:17:16.3321058Z   
2026-02-04T16:17:16.3321294Z   == Results ==
2026-02-04T16:17:16.3321798Z   !== Correct Answer - 3 ==                                   == Spark Answer - 3 ==
2026-02-04T16:17:16.3322827Z    struct<replace(col, , x):string,replace(col, , ):string>   struct<replace(col, , x):string,replace(col, , ):string>
2026-02-04T16:17:16.3323753Z   ![,]                                                        [null,null]
2026-02-04T16:17:16.3324349Z   ![hello,hello]                                              [x,]
2026-02-04T16:17:16.3325067Z   ![null,null]                                                [xhxexlxlxox,hello] (QueryTest.scala:244)

[wattt3]

Hi @andygrove, thanks for the review
Due to environment issues, I couldn't run the test locally at the time.
I changed to a SQL file-based approach, but the result differs from the issue description.