On servers prepared for your deployment scenario:
Install the Postgres Pro AXE DBMS from the distribution or any other supported DBMS.
For more information, refer to the Postgres Pro support team.
Install pgpro_axe and pgpro_metastore:
Install the pgpro_axe package, then the pgpro_metastore package received from Postgres Pro.
In the postgresql.conf configuration file,
set shared_preload_libraries to
pgpro_axe, pgpro_metastore.
Specify the names of extensions in this exact order. Otherwise, they are not installed.
Restart the server.
Create pgpro_axe, then pgpro_metastore in the server database:
CREATE EXTENSION pgpro_axe; CREATE EXTENSION pgpro_metastore;
Ensure that Postgres Pro AXE operates correctly by creating Parquet files in a storage:
Example 12.1. Creating Parquet files in a local or network storage
COPY ( SELECT test_id, test_name FROM tbl_test WHERE test_id between 22 and 30 ) TO '/tmp/data_parquet/test_local.parquet' (FORMAT parquet, COMPRESSION zstd);
Example 12.2. Creating Parquet files in an S3 storage
user@db=# COPY ( SELECT test_id, test_name FROM tbl_test WHERE test_id between 1 and 10 ) TO 's3://axedemobucket/test01.parquet' (FORMAT parquet, COMPRESSION zstd); COPY 10 user@db=# COPY ( SELECT test_id, test_name FROM tbl_test WHERE test_id between 11 and 20 ) TO 's3://axedemobucket/test02.parquet' (FORMAT parquet, COMPRESSION zstd); COPY 10
If Parquet files were created successfully, you can select them:
Example 12.3. Selecting Parquet files from a local or network storage
SELECT r['column_name'] FROM read_parquet('/path/test.parquet') r;
Example 12.4. Selecting Parquet files from an S3 storage
SELECT r['column_name'] FROM read_parquet('s3://bucket_name/test.parquet') r
If you are able to create and select Parquet files, Postgres Pro AXE is installed successfully.