Improve documentation for format OPTIONS clause

datafusion/core/src/lib.rs

@@ -1021,8 +1021,8 @@ doc_comment::doctest!(
 
 #[cfg(doctest)]
 doc_comment::doctest!(
-    "../../../docs/source/user-guide/sql/write_options.md",
-    user_guide_sql_write_options
+    "../../../docs/source/user-guide/sql/format_options.md",
+    user_guide_sql_format_options
 );
 
 #[cfg(doctest)]

docs/source/user-guide/sql/ddl.md

@@ -74,7 +74,7 @@ LOCATION <literal>
 <key_value_list> := (<literal> <literal, <literal> <literal>, ...)
 ```
 
-For a detailed list of write related options which can be passed in the OPTIONS key_value_list, see [Write Options](write_options).
+For a comprehensive list of format-specific options that can be specified in the `OPTIONS` clause, see [Format Options](format_options.md).
 
 `file_type` is one of `CSV`, `ARROW`, `PARQUET`, `AVRO` or `JSON`
 

docs/source/user-guide/sql/dml.md

@@ -49,7 +49,7 @@ The output format is determined by the first match of the following rules:
 1. Value of `STORED AS`
 2. Filename extension (e.g. `foo.parquet` implies `PARQUET` format)
 
-For a detailed list of valid OPTIONS, see [Write Options](write_options).
+For a detailed list of valid OPTIONS, see [Format Options](format_options.md).
 
 ### Examples
 

docs/source/user-guide/sql/format_options.md

@@ -0,0 +1,142 @@
+<!---
+  Licensed to the Apache Software Foundation (ASF) under one
+  or more contributor license agreements.  See the NOTICE file
+  distributed with this work for additional information
+  regarding copyright ownership.  The ASF licenses this file
+  to you under the Apache License, Version 2.0 (the
+  "License"); you may not use this file except in compliance
+  with the License.  You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+  Unless required by applicable law or agreed to in writing,
+  software distributed under the License is distributed on an
+  "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+  KIND, either express or implied.  See the License for the
+  specific language governing permissions and limitations
+  under the License.
+-->
+
+# Format Options
+
+DataFusion supports customizing how data is read from or written to disk as a result of a `COPY`, `INSERT INTO`, or `CREATE EXTERNAL TABLE` query. There are a few special options, file format (e.g., CSV or Parquet) specific options, and Parquet column-specific options. Options can also in some cases be specified in multiple ways with a set order of precedence.
+
+## Specifying Options and Order of Precedence
+
+Format-related options can be specified in the following ways:
+
+- Session-level config defaults
+- `CREATE EXTERNAL TABLE` options
+- `COPY` option tuples
+
+For a list of supported session-level config defaults, see [Configuration Settings](../configs). These defaults apply to all operations but have the lowest level of precedence.
+
+If creating an external table, table-specific format options can be specified when the table is created using the `OPTIONS` clause:
+
+```sql
+CREATE EXTERNAL TABLE
+  my_table(a bigint, b bigint)
+  STORED AS csv
+  LOCATION '/test/location/my_csv_table/'
+  OPTIONS(
+    NULL_VALUE 'NAN',
+    'has_header' 'true',
+    'format.delimiter' ';'
+  )
+```
+
+When running `INSERT INTO my_table ...`, the options from the `CREATE TABLE` will be respected (e.g., gzip compression, special delimiter, and header row included). Note that compression, header, and delimiter settings can also be specified within the `OPTIONS` tuple list. Dedicated syntax within the SQL statement always takes precedence over arbitrary option tuples, so if both are specified, the `OPTIONS` setting will be ignored.
+
+Finally, options can be passed when running a `COPY` command.
+
+```sql
+COPY source_table
+  TO 'test/table_with_options'
+  PARTITIONED BY (column3, column4)
+  OPTIONS (
+    format parquet,
+    compression snappy,
+    'compression::column1' 'zstd(5)',
+  )
+```
+
+In this example, we write the entirety of `source_table` out to a folder of Parquet files. One Parquet file will be written in parallel to the folder for each partition in the query. The next option `compression` set to `snappy` indicates that unless otherwise specified, all columns should use the snappy compression codec. The option `compression::col1` sets an override, so that the column `col1` in the Parquet file will use the ZSTD compression codec with compression level `5`. In general, Parquet options that support column-specific settings can be specified with the syntax `OPTION::COLUMN.NESTED.PATH`.
+
+# Available Options
+
+## Generic Options
+
+| Option     | Description                                                   | Default Value    |
+| ---------- | ------------------------------------------------------------- | ---------------- |
+| NULL_VALUE | Sets the string which should be used to indicate null values. | arrow-rs default |
+
+## Execution-Specific Options
+
+The following options are available when executing a `COPY` query.
+
+| Option                    | Description                                                                        | Default Value |
+| ------------------------- | ---------------------------------------------------------------------------------- | ------------- |
+| KEEP_PARTITION_BY_COLUMNS | Flag to retain the columns in the output data when using `PARTITIONED BY` queries. | false         |
+
+Note: `execution.keep_partition_by_columns` flag can also be enabled through `ExecutionOptions` within `SessionConfig`.
+
+## JSON Format Options
+
+The following options are available when reading or writing JSON files. Note: If any unsupported option is specified, an error will be raised and the query will fail.
+
+| Option      | Description                                                                                                                        | Default Value |
+| ----------- | ---------------------------------------------------------------------------------------------------------------------------------- | ------------- |
+| COMPRESSION | Sets the compression that should be applied to the entire JSON file. Supported values are GZIP, BZIP2, XZ, ZSTD, and UNCOMPRESSED. | UNCOMPRESSED  |
+
+**Example:**
+
+```sql
+CREATE EXTERNAL TABLE t
+STORED AS JSON
+LOCATION '/tmp/foo.json'
+OPTIONS('COMPRESSION', 'gzip');
+```
+
+## CSV Format Options
+
+The following options are available when reading or writing CSV files. Note: If any unsupported option is specified, an error will be raised and the query will fail.
+
+| Option             | Description                                                                                                                       | Default Value    |
+| ------------------ | --------------------------------------------------------------------------------------------------------------------------------- | ---------------- |
+| COMPRESSION        | Sets the compression that should be applied to the entire CSV file. Supported values are GZIP, BZIP2, XZ, ZSTD, and UNCOMPRESSED. | UNCOMPRESSED     |
+| HEADER             | Sets if the CSV file should include column headers                                                                                | false            |
+| NEWLINES_IN_VALUES | Sets if newlines in quoted values are supported                                                                                   | false            |
+| DATE_FORMAT        | Sets the format that dates should be encoded in within the CSV file                                                               | arrow-rs default |
+| DATETIME_FORMAT    | Sets the format that datetimes should be encoded in within the CSV file                                                           | arrow-rs default |
+| TIME_FORMAT        | Sets the format that times should be encoded in within the CSV file                                                               | arrow-rs default |
+| RFC3339            | If true, uses RFC3339 format for date and time encodings                                                                          | arrow-rs default |
+| NULL_VALUE         | Sets the string which should be used to indicate null values within the CSV file.                                                 | arrow-rs default |
+| DELIMITER          | Sets the character which should be used as the column delimiter within the CSV file.                                              | arrow-rs default |
+
+**Example:**
+
+```sql
+CREATE EXTERNAL TABLE t
+STORED AS CSV
+LOCATION '/tmp/foo.csv'
+OPTIONS('DELIMITER', '|', 'HEADER', 'true', 'NEWLINES_IN_VALUES', 'true');
+```
+
+## Parquet Format Options
+
+The following options are available when reading or writing Parquet files. If any unsupported option is specified, an error will be raised and the query will fail. If a column-specific option is specified for a column that does not exist, the option will be ignored without error.
+
+| Option               | Can be Column Specific? | Description                                                                                                                         |
+| -------------------- | ----------------------- | ----------------------------------------------------------------------------------------------------------------------------------- |
+| COMPRESSION          | Yes                     | Sets the compression codec and if applicable compression level to use                                                               |
+| MAX_ROW_GROUP_SIZE   | No                      | Sets the maximum number of rows that can be encoded in a single row group. Larger row groups require more memory to write and read. |
+| BLOOM_FILTER_ENABLED | Yes                     | Sets whether a bloom filter should be written into the file.                                                                        |
+
+**Example:**
+
+```sql
+CREATE EXTERNAL TABLE t
+STORED AS PARQUET
+LOCATION '/tmp/foo.parquet'
+OPTIONS('COMPRESSION', 'snappy', 'MAX_ROW_GROUP_SIZE', '1000000', 'BLOOM_FILTER_ENABLED', 'true');
+```

docs/source/user-guide/sql/index.rst

@@ -33,5 +33,5 @@ SQL Reference
    window_functions
    scalar_functions
    special_functions
-   write_options
+   format_options
    prepared_statements