Databricks - Bulk Load

In this article

Overview

You can use this Snap to perform a bulk load operation on your DLP instance. The source of your data can be a file from a cloud storage location, an input view from an upstream Snap, or a table that can be accessed through a JDBC connection. The source data can be in a CSV, JSON, PARQUET, TEXT, or an ORC file.

This Snap uses the following Databricks commands internally:

  • COPY INTO - Enables loading data from staged files to an existing table.

  • CREATE TABLE [USING] - Enables loading data from some external sources like JDBC.

  • CREATE TABLE - Creates table in our case temporary table.

  • INSERT INTO - Inserts new rows into a table.

Snap Type

Databricks - Bulk Load Snap is a write-type Snap that loads data into your DLP instance.

Prerequisites

  • Valid access credentials to a DLP instance with adequate access permissions to perform the action in context.

  • Valid access to the external source data in one of the following: Azure Blob Storage, ADLS Gen2, DBFS, GCP, AWS S3, or another database (JDBC-compatible).

Support for Ultra Pipelines

Does not support Ultra Pipelines. 

Limitations

  • Snaps in the Databricks Snap Pack do not support array, map, and struct data types in their input and output documents.

  • The Databricks - Bulk Load Snap fails to execute the DROP AND CREATE TABLE and ALTER TABLE operations on tables when using the Databricks SQL persona on the AWS Cloud. The error message Operation not allowed: ALTER TABLE RENAME TO is not allowed for managed Delta tables on S3 is displayed. However, the same actions run successfully when using the Data Science and Engineering persona on the AWS Cloud. This is a limitation on the Databricks endpoint for serverless configurations or SQL endpoint clusters.
    Workaround: If you want to use DROP AND CREATE TABLE action in the Databricks - Bulk Load Snap, then connect a Databricks - Execute Snap upstream of the Bulk Load Snap to drop the table using this syntax: DROP TABLE IF EXISTS <target table name>
    By doing so, the Databricks - Bulk Load Snap does not invoke the ALTER TABLE SQL, and hence the pipeline runs successfully. You can also drop the table through the console before invoking the pipeline that contains the Databricks -Bulk Load Snap.

Known Issues


Cause: This issue arises due to a limitation within the Databricks SQL Admin Console, which prevents you from adding the configuration parameter spark.databricks.delta.alterTable.rename.enabledOnAWS trueto the SQL Warehouse Settings. As a result, the Snap encounters restrictions when attempting to perform certain operations on managed Delta tables stored on Amazon S3.

Snap Views

Type

Format

Number of Views

Examples of Upstream and Downstream Snaps

Description

Type

Format

Number of Views

Examples of Upstream and Downstream Snaps

Description

Input 

Document

  • Min: 0

  • Max: 2

  • Mapper

  • Copy

  • JSON Generator

  • Databricks - Select

This Snap can read from two input documents at a time:

  • One JSON document for the incoming data to be loaded into the target Databricks instance.

  • Another JSON document that contains the table schema (metadata) for creating the target table.

Output

Document

  • Min: 0

  • Max: 1

  • Databricks - Select

  • Databricks - Unload

A JSON document containing the bulk load request details and the result of the bulk load operation.

Error

Error handling is a generic way to handle errors without losing data or failing the Snap execution. You can handle the errors that the Snap might encounter while running the Pipeline by choosing one of the following options from the When errors occur list under the Views tab. The available options are:

  • Stop Pipeline Execution: Stops the current pipeline execution when the Snap encounters an error.

  • Discard Error Data and Continue: Ignores the error, discards that record, and continues with the rest of the records.

  • Route Error Data to Error View: Routes the error data to an error view without stopping the Snap execution.

Learn more about Error handling in Pipelines.

Snap Settings

  • Asterisk ( * ): Indicates a mandatory field.

  • Suggestion icon (): Indicates a list that is dynamically populated based on the configuration.

  • Expression icon ( ): Indicates whether the value is an expression (if enabled) or a static value (if disabled). Learn more about Using Expressions in SnapLogic.

  • Add icon ( ): Indicates that you can add fields in the fieldset.

  • Remove icon ( ): Indicates that you can remove fields from the fieldset.

Field Name

Field Type

Field Dependency

Description

Field Name

Field Type

Field Dependency

Description

Label*

 

Default Value: Databricks - Bulk Load
Example: Db_BulkLoad_FromS3

String

None

The name for the Snap. You can modify this to be more specific, especially if you have more than one of the same Snap in your Pipeline.

Database name

 

Default Value: None.
Example: cust_db

String/Expression/Suggestion

None

Enter the name of the database in which the target table exists. Leave this blank if you want to use the database name specified in the Database Name field in the account settings.

Table Name*

 

Default Value: None.
Example: cust_records

String/Expression/Suggestion

None

Enter the name of the table in which you want to perform the bulk load operation. 

Source Type

 

Default Value: Cloud Storage File
Example: Input View

Dropdown list

None

Select the type of source from which you want to load the data into your DLP instance. The available options are:

  • Cloud Storage File. A file from a cloud location like AWS S3, Azure, or GCS. You can configure a series of options for the bulk load operation as described in this document.

  • Input View. A JSON file coming from the preceding Snap’s output. You need to specify only the Load action.

  • JDBC. A table in another database that can be connected to using a JDBC connector. You can specify the Source table name to load the data from or the Target Table Columns to replace the existing target table with a new one.

Load action*

 

Default Value: Drop and create table
Example: Append rows to existing table

Dropdown list

None

Select the appropriate load action you want to perform on the target table for this bulk upload operation. You can:

  • Drop and create a table. To remove existing table in the specified database and create a new table with the schema defined below in the Snap, from an Input View, or a JDBC-connected database table.

  • Append rows to existing table. To insert new rows of data to an existing target table.

Source table name

String

Source Type is JDBC.

Enter the source table name. The default values (database) configured in the Snap’s account for JDBC Account type are considered, if not specified in this field.

Target Table Columns

 

Source Type is Cloud Storage file or JDBC and Load action is Drop and create table.

Use this fieldset to specify the target table schema for creating a new table. Specify the Column Name and Data Type for as many columns you need to load in the target table.

Column

Default Value: None.
Example: cust_ID

String

None

Enter the name of the column that you want to load in the target table.

Data Type

Default Value: None.
Example: int, string

String

None

Enter the data type of the values in the specified column.

File format type

 

Default Value: CSV
Example: PARQUET

Dropdown list

Source Type is Cloud Storage file.

Select the file format of the source data file. It can be CSV, JSON, ORC, PARQUET, or TEXT.

File Format Option List

 

Source Type is Cloud Storage file.

You can use this field set to choose the file format options to associate with the bulk load operation, based on your source file format. Choose one file format option in each row.

File format option

 

Default Value: None.
Example: cust_ID

String/Expression/Suggestion

Source Type is Cloud Storage file.

Select a file format option from the available options and set appropriate values to suit your bulk load needs, without affecting the syntax displayed in this field.

Files provider

 

Default Value: File list
Example: pattern

Dropdown list

Source Type is Cloud Storage file.

Declare the manner in which you are specifying the source files list - File list or pattern. Based on your selection in this field, the corresponding fields change: File list fieldset for File list and File pattern field for pattern.

File list

 

Source Type is Cloud Storage file and Files provider is File list.

You can use this field set to specify the file paths to be used for the bulk load operation. Choose one file path in each row.

File

 

Default Value: None.
Example: cust_data.csv

String

Source Type is Cloud Storage file and Files provider is File list.

Enter the path of the file to be used for the bulk upload operation.

File pattern

 

Default Value: None.
Example: folder1/file_[a-g].csv

String/Expression

Source Type is Cloud Storage file and Files provider is pattern.

Enter the regex pattern to use to match the file name and/or absolute path. You can specify this as a regular expression pattern string, enclosed in single quotes. Learn more: Examples of COPY INTO (Delta Lake on Databricks) for DLP.

Encryption type

 

Default Value: None.
Example: Server-Side KMS Encryption

String

Source Type is Cloud Storage file.

Select the encryption type to use for decrypting the source data and/or files staged in the S3 buckets.

Server-side encryption is available only for S3 accounts.

KMS key

 

Default Value: None.
Example: MF96D-M9N47-XKV7X-C3GCQ-G5349

String/Expression

Source Type is Cloud Storage file and Encryption type is Server-Side KMS Encryption.

Enter the KMS key to use to encrypt the files. In case that your source files are in S3, see Loading encrypted files from Amazon S3 for more detail.

Number of Retries

Example: 3

Minimum value: 0

Default value: 0

Integer

Source Type is Input View.

Specifies the maximum number of retry attempts when the Snap fails to write.

 

Retry Interval (seconds)

Example: 3

Minimum value: 1

Default value: 1

Integer

Source Type is Input View.

Specifies the minimum number of seconds the Snap must wait before each retry attempt.

Manage Queued Queries

Default value: Continue to execute queued queries when pipeline is stopped or if it fails.

Example: Cancel queued queries when pipeline is stopped or if it fails

Dropdown list

None

Select this property to determine whether the Snap should continue or cancel the execution of the queued Databricks SQL queries when you stop the Pipeline.

If you select Cancel queued queries when pipeline is stopped or if it fails, then the read queries under execution are cancelled, whereas the write type of queries under execution are not cancelled. Databricks internally determines which queries are safe to be cancelled and cancels those queries.

Snap Execution

 

Default Value: Execute only
Example: Validate & Execute

Dropdown list

None

Select one of the three modes in which the Snap executes. Available options are:

  • Validate & Execute: Performs limited execution of the Snap, and generates a data preview during Pipeline validation. Subsequently, performs full execution of the Snap (unlimited records) during Pipeline runtime.

  • Execute only: Performs full execution of the Snap during Pipeline execution without generating preview data.

  • Disabled: Disables the Snap and all Snaps that are downstream from it.

Troubleshooting

Error

Reason

Resolution

Error

Reason

Resolution

Missing property value

You have not specified a value for the required field where this message appears.

Ensure that you specify valid values for all required fields.

Examples

Bulk Load Employee data from a CSV file into a DLP instance

Consider the scenario where we need the employee data from a CSV file to be fed into a DLP instance so that we can analyze the data.

Prerequisite:

  • Configure the Bulk Load Snap account to connect to the AWS S3 service using Source Location Credentials to read the CSV file.

  • We need two Snaps:

    • Databricks Bulk Load: To load the data from the CSV file in an S3 location

    • Databricks Select: To read the data loaded in the target table and generate some insights.

Configure the Databricks - Bulk Load Snap to load employee data from the CSV into a new table, company_employees.

Here is how we do it:

  • Select the Drop and create table as the Load action.

  • Define the schema for the new table in the Target Table Columns field set.

  • Choose the source data type and indicate that the file contains a valid header.

  • Specify the file names (with relative paths, here) to load the data from.

  • As our CSV file in the S3 location is not encrypted, we leave the corresponding fields blank.

Run the pipeline—it loads the valid data into the target table and displays the new table name and the number of records loaded.

Next, to read the data from the new table in the DLP instance, use the Databricks - Select Snap. Provide the Table name and configure the Snap with a WHERE clause, salary < 500000.

On validation, the Snap retrieves and displays the data from the company_employees table that matches the WHERE condition specified.

Download this Pipeline. 

Downloads

 

  File Modified

File Databricks_FEP1_BulkLoad.slp

Jul 19, 2022 by Anand Vedam

Snap Pack History


Related Links