[SPARK-38358][DOC] Add migration guide for spark.sql.hive.convertMetastoreInsertDir, spark.sql.hive.convertInsertingPartitionedTable and spark.sql.hive.convertMetastoreCtas

docs/sql-migration-guide.md

@@ -62,6 +62,8 @@ license: |
 
   - Since Spark 3.3, when reading values from a JSON attribute defined as `FloatType` or `DoubleType`, the strings `"+Infinity"`, `"+INF"`, and `"-INF"` are now parsed to the appropriate values, in addition to the already supported `"Infinity"` and `"-Infinity"` variations. This change was made to improve consistency with Jackson's parsing of the unquoted versions of these values. Also, the `allowNonNumericNumbers` option is now respected so these strings will now be considered invalid if this option is disabled.
 
+  - Since Spark 3.3, Spark will try to use built-in data source writer instead of Hive serde in `INSERT OVERWRITE DIRECTORY`. This behavior is effective only if `spark.sql.hive.convertMetastoreParquet` or `spark.sql.hive.convertMetastoreOrc` is enabled respectively for Parquet and ORC formats. To restore the behavior before Spark 3.3, you can set `spark.sql.hive.convertMetastoreInsertDir` to `false`.
+
 ## Upgrading from Spark SQL 3.1 to 3.2
 
   - Since Spark 3.2, ADD FILE/JAR/ARCHIVE commands require each path to be enclosed by `"` or `'` if the path contains whitespaces.
@@ -352,6 +354,10 @@ license: |
 
   - In Spark 3.0, datetime pattern letter `F` is **aligned day of week in month** that represents the concept of the count of days within the period of a week where the weeks are aligned to the start of the month. In Spark version 2.4 and earlier, it is **week of month** that represents the concept of the count of weeks within the month where weeks start on a fixed day-of-week, e.g. `2020-07-30` is 30 days (4 weeks and 2 days) after the first day of the month, so `date_format(date '2020-07-30', 'F')` returns 2 in Spark 3.0, but as a week count in Spark 2.x, it returns 5 because it locates in the 5th week of July 2020, where week one is 2020-07-01 to 07-04.
 
+  - In Spark 3.0, Spark will try to use built-in data source writer instead of Hive serde in `CTAS`. This behavior is effective only if `spark.sql.hive.convertMetastoreParquet` or `spark.sql.hive.convertMetastoreOrc` is enabled respectively for Parquet and ORC formats. To restore the behavior before Spark 3.0, you can set `spark.sql.hive.convertMetastoreCtas` to `false`.
+
+  - In Spark 3.0, Spark will try to use built-in data source writer instead of Hive serde to process inserting into partitioned ORC/Parquet tables created by using the HiveSQL syntax. This behavior is effective only if `spark.sql.hive.convertMetastoreParquet` or `spark.sql.hive.convertMetastoreOrc` is enabled respectively for Parquet and ORC formats. To restore the behavior before Spark 3.0, you can set `spark.sql.hive.convertInsertingPartitionedTable` to `false`.
+
 ### Data Sources
 
   - In Spark version 2.4 and below, when reading a Hive SerDe table with Spark native data sources(parquet/orc), Spark infers the actual file schema and update the table schema in metastore. In Spark 3.0, Spark doesn't infer the schema anymore. This should not cause any problems to end users, but if it does, set `spark.sql.hive.caseSensitiveInferenceMode` to `INFER_AND_SAVE`.