What causes TABLE_ALREADY_EXISTS?

A job was re-run and attempted to recreate a table without dropping it first. An idempotent CREATE TABLE IF NOT EXISTS clause is missing from the DDL. A Spark DataFrame saveAsTable call used SaveMode.ErrorIfExists (the default) on an existing table. A previous failed run partially created the table before failing

How do I fix TABLE_ALREADY_EXISTS?

Step 1: Add IF NOT EXISTS to the CREATE TABLE statement: CREATE TABLE IF NOT EXISTS catalog.schema.table .... Step 2: For DataFrame writes, use mode('overwrite') or mode('append') instead of the default ErrorIfExists: df.write.mode('overwrite').saveAsTable('...'). Step 3: If the existing table is stale or incorrect, drop it first with DROP TABLE IF EXISTS.. Step 4: Review job logic to ensure idempotency — a robust pipeline should be able to re-run without failures.

Medium severityexecution

Power BI Refresh Error:
TABLE_ALREADY_EXISTS

What does this error mean?

A CREATE TABLE statement failed because a table with the same name already exists in the target schema.

Common causes

1A job was re-run and attempted to recreate a table without dropping it first
2An idempotent CREATE TABLE IF NOT EXISTS clause is missing from the DDL
3A Spark DataFrame saveAsTable call used SaveMode.ErrorIfExists (the default) on an existing table
4A previous failed run partially created the table before failing

How to fix it

1Step 1: Add IF NOT EXISTS to the CREATE TABLE statement: CREATE TABLE IF NOT EXISTS catalog.schema.table ...
2Step 2: For DataFrame writes, use mode('overwrite') or mode('append') instead of the default ErrorIfExists: df.write.mode('overwrite').saveAsTable('...')
3Step 3: If the existing table is stale or incorrect, drop it first with DROP TABLE IF EXISTS.
4Step 4: Review job logic to ensure idempotency — a robust pipeline should be able to re-run without failures.

Frequently asked questions

What is the Spark equivalent of CREATE TABLE IF NOT EXISTS for DataFrame writes?

Use df.write.mode('ignore').saveAsTable('table_name') to skip the write if the table already exists, or mode('overwrite') to replace it.

Can I use REPLACE TABLE instead of DROP + CREATE?

Yes — CREATE OR REPLACE TABLE is supported in Databricks SQL and is atomic: it replaces the existing table definition and data in a single operation.