Question 1

Why export lakehouse data into a bucket?

Accepted Answer

The reasons vary, but they generally come down to something needing files rather than a table it can query. Handing data to a team or partner working in another environment, feeding a processing job, or keeping a record on your own retention terms are all common. Since this runs on a schedule rather than as a manual extract, the bucket stays current instead of representing whenever someone last exported.

Question 2

Which format preserves the data best?

Accepted Answer

Parquet, for anything read by a query engine afterwards, because it carries types and reads efficiently by column. That matters here because Databricks exposes STRUCT, ARRAY and MAP, and a format that cannot represent nesting will flatten or stringify them. JSON and JSON Lines also keep nested values intact and suit record by record processing. CSV remains right where the consumer expects it, at the cost of type information.

Question 3

How do I organise the output?

Accepted Answer

Through expressions rather than a partitioning setting. The path is the bucket and prefix on the connection, then an optional subfolder, then the output name, and both the subfolder and name are expression evaluated, so a datetime expression produces dated folders as runs happen. This is path templating rather than partitioning on a value in the data, and there is no compression codec setting to configure.

Question 4

What drives runtime on the Databricks side?

Accepted Answer

The SQL warehouse the connection uses. Reads go through it, so its size, autoscaling behaviour and current concurrency all affect how long an export takes, alongside how much data the query scans. A cold warehouse adds start up time before anything moves, with the connection retried for around ten minutes while it comes up, and serverless warehouses generally start faster than classic ones.

Question 5

How do I sync Databricks to Google Cloud Storage?

Accepted Answer

Connect Databricks with a SQL warehouse hostname and a token, or an OAuth service principal, and connect Cloud Storage with a service account JSON key and the destination bucket. Then select the data to export, choose CSV, JSON Lines, JSON array or Parquet, and pick the write mode. Parquet is the usual choice where the files will be read by a query engine afterwards.

Connect Databricks with Google Cloud Storage

Connect Databricks and Google Cloud Storage with Polytomic

Four output formats

Service account on the bucket

Snapshots or a stable file

Trusted by businesses worldwide

Step-by-step integration guide

Connect Databricks in Polytomic

Connect Google Cloud Storage in Polytomic

Create a sync between Databricks and Google Cloud Storage

Map fields and enable the sync

Frequently asked questions

The sync and ETL platform that does it all

Powerful data sync

Advanced analytics

Enterprise security

Easy setup

Ready to connect Databricks with Google Cloud Storage?