BigQuery

BigQuery is a completely serverless and cost-effective enterprise data warehouse that works across clouds and scales with your data, with BI, machine learning and AI built in.

The BigQuery Wrapper allows you to read and write data from BigQuery within your Postgres database.

Preparation#

Before you can query BigQuery, you need to enable the Wrappers extension and store your credentials in Postgres.

Enable Wrappers#

Make sure the wrappers extension is installed on your database:

1
create extension if not exists wrappers with schema extensions;

Enable the BigQuery Wrapper#

Enable the bigquery_wrapper FDW:

1
2
3
create foreign data wrapper bigquery_wrapper  handler big_query_fdw_handler  validator big_query_fdw_validator;

Store your credentials (optional)#

By default, Postgres stores FDW credentials inside pg_catalog.pg_foreign_server in plain text. Anyone with access to this table will be able to view these credentials. Wrappers is designed to work with Vault, which provides an additional level of security for storing credentials. We recommend using Vault to store your credentials.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
-- Save your BigQuery service account json in Vault and retrieve the `key_id`insert into vault.secrets (name, secret)values (  'bigquery',  '    {      "type": "service_account",      "project_id": "your_gcp_project_id",      "private_key_id": "your_private_key_id",      "private_key": "-----BEGIN PRIVATE KEY-----\n...\n-----END PRIVATE KEY-----\n",      ...    }  ')returning key_id;

Connecting to BigQuery#

We need to provide Postgres with the credentials to connect to BigQuery, and any additional options. We can do this using the create server command:

1
2
3
4
5
6
7
create server bigquery_server  foreign data wrapper bigquery_wrapper  options (    sa_key_id '<key_ID>', -- The Key ID from above.    project_id 'your_gcp_project_id',    dataset_id 'your_gcp_dataset_id'  );

Create a schema#

We recommend creating a schema to hold all the foreign tables:

1
create schema if not exists bigquery;

Options#

The following options are available when creating BigQuery foreign tables:

• table - Source table or view name in BigQuery, required
• location - Source table location (default: 'US')
• timeout - Query request timeout in milliseconds (default: 30000)
• rowid_column - Primary key column name (required for data modification)

You can also use a subquery as the table option:

1
table '(select * except(props), to_json_string(props) as props from `my_project.my_dataset.my_table`)'

Note: When using subquery, full qualified table name must be used.

Entites#

Tables#

The BigQuery Wrapper supports data reads and writes from BigQuery tables and views.

Operations#

Usage#

1
2
3
4
5
6
7
8
9
10
create foreign table bigquery.my_bigquery_table (  id bigint,  name text,  ts timestamp)  server bigquery_server  options (    table 'people',    location 'EU'  );

Notes#

• Supports where, order by and limit clause pushdown
• When using rowid_column, it must be specified for data modification operations
• Data in the streaming buffer cannot be updated or deleted until the buffer is flushed (up to 90 minutes)

Query Pushdown Support#

This FDW supports where, order by and limit clause pushdown.

Inserting Rows & the Streaming Buffer#

This foreign data wrapper uses BigQuery’s insertAll API method to create a streamingBuffer with an associated partition time. Within that partition time, the data cannot be updated, deleted, or fully exported. Only after the time has elapsed (up to 90 minutes according to BigQuery’s documentation), can you perform operations.

If you attempt an UPDATE or DELETE statement on rows while in the streamingBuffer, you will get an error of UPDATE or DELETE statement over table datasetName - note that tableName would affect rows in the streaming buffer, which is not supported.

Supported Data Types#

Limitations#

This section describes important limitations and considerations when using this FDW:

• Large result sets may experience network latency during data transfer
• Data in streaming buffer cannot be modified for up to 90 minutes
• Only supports specific data type mappings between Postgres and BigQuery
• Materialized views using foreign tables may fail during logical backups

Examples#

Some examples on how to use BigQuery foreign tables.

Let's prepare the source table in BigQuery first:

1
2
3
4
5
6
7
8
9
10
11
12
13
-- Run below SQLs on BigQuery to create source tablecreate table your_project_id.your_dataset_id.people (  id int64,  name string,  ts timestamp,  props jsonb);-- Add some test datainsert into your_project_id.your_dataset_id.people values  (1, 'Luke Skywalker', current_timestamp(), parse_json('{"coordinates":[10,20],"id":1}')),  (2, 'Leia Organa', current_timestamp(), null),  (3, 'Han Solo', current_timestamp(), null);

Basic example#

This example will create a "foreign table" inside your Postgres database called people and query its data:

1
2
3
4
5
6
7
8
9
10
11
12
13
create foreign table bigquery.people (  id bigint,  name text,  ts timestamp,  props jsonb)  server bigquery_server  options (    table 'people',    location 'EU'  );select * from bigquery.people;

Data modify example#

This example will modify data in a "foreign table" inside your Postgres database called people, note that rowid_column option is mandatory:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
create foreign table bigquery.people (  id bigint,  name text,  ts timestamp,  props jsonb)  server bigquery_server  options (    table 'people',    location 'EU',    rowid_column 'id'  );-- insert new datainsert into bigquery.people(id, name, ts, props)values (4, 'Yoda', '2023-01-01 12:34:56', '{"coordinates":[10,20],"id":1}'::jsonb);-- update existing dataupdate bigquery.peopleset name = 'Anakin Skywalker', props = '{"coordinates":[30,40],"id":42}'::jsonbwhere id = 1;-- delete datadelete from bigquery.peoplewhere id = 2;