Moving to drop table instead of archive (#1074)

## Summary

Since BQMS does not support table rename, dropping table instead of
archiving it.

## Checklist
- [ ] Added Unit Tests
- [x] Covered by existing CI
- [ ] Integration tested
- [ ] Documentation update



<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->

## Summary by CodeRabbit

- Behavior Changes
- On schema changes detected during join output, the system now drops
the existing table before saving new data, rather than archiving it.
- Ensures the latest schema is applied immediately on reload, without
retaining archived copies.
- No changes to user-facing flows; compute, save, and reload continue as
before.
  - Log messages updated to reflect the drop-on-schema-change action.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->

Co-authored-by: ezvz <[email protected]>

Author: varant-zlai

spark/src/main/scala/ai/chronon/spark/JoinBase.scala

@@ -241,7 +241,7 @@ abstract class JoinBase(val joinConfCloned: api.Join,
     dfOpt.map { df =>
       val table = joinMetaData.outputTable
 
-      tableUtils.archiveTableOnSchemaChange(table, df)
+      tableUtils.dropTableOnSchemaChange(table, df)
       df.save(table)
 
       tableUtils.loadTable(table, range.whereClauses)

spark/src/main/scala/ai/chronon/spark/catalog/TableUtils.scala

@@ -476,7 +476,7 @@ class TableUtils(@transient val sparkSession: SparkSession) extends Serializable
     }
   }
 
-  def archiveTableOnSchemaChange(tableName: String, incomingDf: DataFrame): Unit = {
+  def dropTableOnSchemaChange(tableName: String, incomingDf: DataFrame): Unit = {
     if (!tableReachable(tableName)) return
 
     val existingSchema = loadTable(tableName).schema
@@ -527,7 +527,7 @@ class TableUtils(@transient val sparkSession: SparkSession) extends Serializable
          |""".stripMargin)
 
     if (addedCols.nonEmpty || removedCols.nonEmpty || updatedCols.nonEmpty) {
-      archiveTableIfExists(tableName, None)
+      dropTableIfExists(tableName)
     }
 
   }