Postgres DB

Guide to Integrate your Postgres database with Sprinkle

Datasource Concepts

Before setting up the data import, learn about data import concepts here

Step-by-Step Guide

STEP-1: Allow Postgres to accept connection from Sprinkle

1. Network Connectivity:

   • If Postgres server is on public network, accessible over public IP, allow inbound connection on postgres port (default is 5432) from Sprinkle IPs (34.93.254.126, 34.93.106.136)

   • If Postgres server is on private network, configure SSH Tunnel in Advanced Settings.

2. Create a Read-Only user, providing any name like "sprinkle"

STEP-2: Configure Postgres Connection

To learn about Connection, refer here

• Log into the Sprinkle application

• Navigate to Ingest -> Connections Tab -> Setup Connections ->

• Select Postgres database

  • Provide all the mandatory details

    • Distinct Name: Name to identify this connection

    • Host: Provide IP address or Host name.

    • Port: Provide the Port number.

    • Database: Provide database name if there is any, it should be an existing database.

    • Username

    • Password

    • Advanced Settings: If Yes:-

      • Connection Properties: You can provide optional connection properties. ex- key1=value1&key2=value2

      • Timezone: The timezone selected will be used to fetch Datetime/Timestamp column value.

      • Change Tracking: If enabled, Postgres binlog is used to fetch changed/new rows. This setting cannot be changed later once selected. For more details see this.

      • Connect via SSH Host: If Yes:-

        • SSH Host: IP address or hostname of the SSH server.

        • SSH Public Key: Add this public key to the ~/.ssh/authorized_keys file on the ssh host machine.

        • SSH Login Username

• Test Connection

• Create

STEP-3: Configure Postgres data import

To learn about data import, refer here

• Navigate to Ingest -> Data ImportsTab -> Setup Sources ->

• Select Postgres database

• Provide the name -> Create

• Connection Tab:

  • From the drop-down, select the name of the connection created in STEP-2

  • Update

STEP-4: Create a Dataset

Datasets Tab: To learn about Dataset, refer here.

Add a Dataset for each table that you want to replicate, providing the following details

• Database Name (Required): Database name to select the table from

• Table Type: (Required) :

  • Table: Select Table if you want to replicate the table

    • Source Table (Required): Name of the Postgres table

    • Exclude Columns (Optional): List of columns to exclude

    • Mask Columns (Optional): List of columns to be masked using SHA256

    • Mode (Required):

      • Complete: Ingest full data from the source table in every ingestion job run. Choose this option if your table size is small (<1 million rows) and you want to ingest it infrequently (a few times a day)

      • Incremental: Ingest only the changed or inserted rows in every ingestion job run. Choose this option if your table size is large and you want to ingest in real-time mode. This requires a time column, and a unique key to be present in the table.

        • Remove Duplicate Rows: Duplicate rows are removed by taking the latest value of your unique keys

          • Unique Keys: List of columns forming the unique id

          • Fetch With Delay(In minutes): Use this to apply delay while downloading records. When a non-zero value is specified, records inserted/updated on the source before given minutes will be ingested, instead of data till the current time. Useful when current data is not fully reflected on the source due to heavy loads. It can be left at 0 minutes if not needed.

        • Time Column: Rows with newer time value column will be ingested in each run

        • Destination Table Partition: It will create partitions of the destination table according to the specified column (for Eg: Date). Its value is NO by default. If YES, destination table will be partitioned, else normal materialized table will be created. This feature is only applicable in BigQuery.

          • Partition Column: Enter the partition column name. Partition will be created on the destination table using partition column value. Selected column values should be static even if other fields are updated. This should be of Date type or easily converted into Date format. If no column is selected, Partition will be created on the basis of job run time.

          • Partition Date Format: Enter the format of date.

        • Periodic Full Fetch: If yes, Periodically full table will be fetched from source. Would be useful if data gets deleted from source and you want to keep the data in sync with source. If no, only incremental data will be pulled in every run.

        To Know more about Ingestion Modes, refer here

    • Filter clause: Filter the data while ingesting. (select * from table_name where CONDITION) only give CONDITION here.

  • Query: Select Query, if you want to insert the output of the Postgres query into the data warehouse

    • Query(SQL) (Required): Query that will be executed on Postgres in each ingestion job run. If schema is not selected as default one, provide fully qualified name(mention schema_name in query). Ex- select * from schema_name.table_name;

• Destination Schema (Required) : Data warehouse schema where the table will be ingested into

• Destination Table name (Optional) : It is the table name to be created on the warehouse. If not given, sprinkle will create like ds_<datasourcename>_<tablename>

• Destination Create Table Clause: Provide additional clauses to warehouse-create table queries such as clustering, partitioning, and more, useful for optimizing DML statements. Learn more on how to use this field.

• Create

STEP-5: Run and schedule Ingestion

In the Ingestion Jobs tab:

• Trigger the Job, using Run button

• To schedule, enable Auto-Run. Change the frequency if needed