Azure Databricks - bulk insert to Azure SQL

Recently I had request to load the processed data into the azure SQL database from databricks. databricks program processing around 300 to 400 million records and aggregating those records into certain buckets. Even after the aggregation total number of records going inside the azure SQL database is 40 million. we found that the insertion is happening raw by raw and hence thought of doing the same using bulk insert option provided by the databricks.

databricks provided super documentation on bulk insert and I just followed the same.

to achieve this we need to create Spark connector library which can be done by using the upload option from the cluster.

Click on the library option and provide the coordinate and create the library as mentioned in the below figure.

once the library is created we used below code to execute the bulk insert. database name, user name, password, table name mentioned here are only for illustration purpose only. we had total 25 columns. we can either provide the metadata here or leave it blank but it is recommended to provide as it will improve the performance. if the metadata is not provided, then databricks match the target system metadata before the actual bulk load.

val deftable = spark.sql("select column1,column2 from tablename")
var bulkCopyMetadata = new BulkCopyMetadata
bulkCopyMetadata.addColumnMetadata(1, "ID", java.sql.Types.VARCHAR, 100, 0)
bulkCopyMetadata.addColumnMetadata(2, "Name", java.sql.Types.VARCHAR, 1000, 0)

val bulkCopyConfig = Config(Map(
  "url"               -> "",
  "databaseName"      -> "databasename",
  "user"              -> "Username",
  "password"          -> "Password",
  "databaseName"      -> "databasename",
  "dbTable"           -> "dbo.tabletempbulk",
  "bulkCopyBatchSize" -> "1000000",
  "bulkCopyTableLock" -> "true",
  "bulkCopyTimeout"   -> "6000"
deftable.bulkCopyToSqlDB(bulkCopyConfig, bulkCopyMetadata)

Ref :-


  1. Excellent blog since I have visited is really awesome. The important thing is that in this blog content written clearly and understandable. The content of information is very informative. We are also providing the best services click on below links to visit our website.
    Snowflake Training
    Workday Training
    Okta Training
    AEM Training
    CyberArk Training

  2. The above wont work with Databricks Runtime 7.3 and above. Any suggestions to run the above code on these DBRs ?


Post a Comment

Popular posts from this blog

Microsoft BI Implementation - Cube back up and restore using XMLA command

Databricks - incorrect header check