INSERT OVERWRITE DIRECTORY
Description
The INSERT OVERWRITE DIRECTORY
statement overwrites the existing data in the directory with the new values using either spark file format or Hive Serde.
Hive support must be enabled to use Hive Serde. The inserted rows can be specified by value expressions or result from a query.
Syntax
INSERT OVERWRITE [ LOCAL ] DIRECTORY [ directory_path ]
{ spark_format | hive_format }
{ VALUES ( { value | NULL } [ , ... ] ) [ , ( ... ) ] | query }
While spark_format
is defined as
USING file_format [ OPTIONS ( key = val [ , ... ] ) ]
hive_format
is defined as
[ ROW FORMAT row_format ] [ STORED AS hive_serde ]
Parameters
-
directory_path
Specifies the destination directory. The
LOCAL
keyword is used to specify that the directory is on the local file system. In spark file format, it can also be specified inOPTIONS
usingpath
, butdirectory_path
andpath
option can not be both specified. -
file_format
Specifies the file format to use for the insert. Valid options are
TEXT
,CSV
,JSON
,JDBC
,PARQUET
,ORC
,HIVE
,LIBSVM
, or a fully qualified class name of a custom implementation oforg.apache.spark.sql.execution.datasources.FileFormat
. -
OPTIONS ( key = val [ , … ] )
Specifies one or more options for the writing of the file format.
-
hive_format
Specifies the file format to use for the insert. Both
row_format
andhive_serde
are optional.ROW FORMAT SERDE
can only be used withTEXTFILE
,SEQUENCEFILE
, orRCFILE
, whileROW FORMAT DELIMITED
can only be used withTEXTFILE
. If both are not defined, spark usesTEXTFILE
. -
row_format
Specifies the row format for this insert. Valid options are
SERDE
clause andDELIMITED
clause.SERDE
clause can be used to specify a customSerDe
for this insert. Alternatively,DELIMITED
clause can be used to specify the nativeSerDe
and state the delimiter, escape character, null character, and so on. -
hive_serde
Specifies the file format for this insert. Valid options are
TEXTFILE
,SEQUENCEFILE
,RCFILE
,ORC
,PARQUET
, andAVRO
. You can also specify your own input and output format usingINPUTFORMAT
andOUTPUTFORMAT
. -
VALUES ( { value
|
NULL } [ , … ] ) [ , ( … ) ]Specifies the values to be inserted. Either an explicitly specified value or a NULL can be inserted. A comma must be used to separate each value in the clause. More than one set of values can be specified to insert multiple rows.
-
query
A query that produces the rows to be inserted. It can be in one of following formats:
- a SELECT statement
- a Inline Table statement
- a
FROM
statement
Examples
Spark format
INSERT OVERWRITE DIRECTORY '/tmp/destination'
USING parquet
OPTIONS (col1 1, col2 2, col3 'test')
SELECT * FROM test_table;
INSERT OVERWRITE DIRECTORY
USING parquet
OPTIONS ('path' '/tmp/destination', col1 1, col2 2, col3 'test')
SELECT * FROM test_table;
Hive format
INSERT OVERWRITE LOCAL DIRECTORY '/tmp/destination'
STORED AS orc
SELECT * FROM test_table;
INSERT OVERWRITE LOCAL DIRECTORY '/tmp/destination'
ROW FORMAT DELIMITED FIELDS TERMINATED BY ','
SELECT * FROM test_table;