Google BigQuery connector (DataStage)

Use the Google BigQuery connector in DataStage® to write data to a BigQuery data warehouse

Prerequisite

Create the connection. For instructions, see Connecting to a data source in DataStage and the Google BigQuery connection.

Configure the Google BigQuery connector for write mode

Specify how the Google BigQuery connector writes data to a BigQuery table. Select the Write mode on the target connector under General target properties. The following table shows the ways that you can configure the connector for the Write mode.

Table 1. Google BigQuery connector write mode
Method	Description
Insert	The Insert mode inserts data into the specified table by using the BigQuery streaming API. Select this method if you want to query data in the table for real-time analysis. It can take up to 90 minutes for the data in the table to be available for subsequent operations like Update, Replace, or Truncate. Meanwhile, if you do any of these operations, you might receive error messages because these operations can cause inconsistencies in table data. For more information, see Streaming data availability in the Google Cloud documentation.
Update then Insert	Update then Insert mode first updates the table based on the key columns that you specify, and then inserts the data into the table. You can also use this mode for Insert operations by not specifying a key column. The Update then Insert mode is the best mode to insert large amounts of data because it provides better performance and helps you avoid streaming errors. With this method, you can do all the operations on the table immediately.
Update	Update mode updates only the table data based on the key columns.
Delete	Delete mode deletes only the table data based on the key columns.
Delete then Insert	Delete then Insert mode first deletes the table data based on the key columns and the inserts the data into the table.

Configure the Google BigQuery connector table action

Specify the action to take on the target table to handle the new data set. The following table shows the ways that you can configure the Table action.

Table 2. Google BigQuery connector table actions
Method	Description
Append	Append creates the table if the table doesn’t exist. If the table already exists, no action is performed.
Replace	Replace drops the existing table and creates a new table with the existing job design schema. An existing table is dropped only when there is no existing streaming buffer attached to the table.
Truncate	Truncate deletes all the records from a table, but it does not remove the table itself.

Partitioned reads

You can configure the Google BigQuery to run on multiple processing nodes and read data from the data source. Each of the processing nodes for the stage retrieves a set (partition) of records from the data source. The partitions are then combined to produce the entire result set for the output link of the stage.

To enable partitioned reads, in the Source stage under General source properties, for Read method, select Select statement. Then select Enable partitioned reads.

When the connector is configured for partitioned reads, it runs the statement that is specified in the Select statement property on each processing node. You can use special placeholders in the statements to ensure that each of the processing nodes retrieves a distinct partition of records from the data source.

You can use the following placeholders in the statements to specify distinct partitions of records to retrieve in individual processing nodes:

[[node-number]]: The node number is replaced at run time with the index of the current processing node. The indexes are zero-based. The placeholder is replaced with the value 0 on the first processing node, value 1 on the second processing node, value 2 on the third processing node, and so forth.
[[node-count]]: The total number of processing nodes for the stage. By default, this number is the number of nodes in the parallel configuration file. The location of the parallel configuration file is specified in the APT_CONFIG_FILE environment variable.

Example SQL statements:

select * from testds.testPartitionRead  where MOD(C1, [[node-count]]) = [[node-number]]

select * from testds.testPartitionRead  where MOD(ABS(FARM_FINGERPRINT(C2)), [[node-count]]) = [[node-number]]