Docs: Fix MERGE INTO example in Getting Started

[varun-lakhyani]

Description

This resolves the TODO noted in spark/v3.5/spark-runtime/src/integration/java/org/apache/iceberg/spark/TestRoundTrip.java:41.

Fixed broken MERGE INTO example in Spark Getting Started documentation that referenced a non-existent count column.

What was wrong

The original example attempted to update a count column:

MERGE INTO local.db.target t USING (SELECT * FROM updates) u ON t.id = u.id
WHEN MATCHED THEN UPDATE SET t.count = t.count + u.count
WHEN NOT MATCHED THEN INSERT *;

However, the table schema defined earlier in the guide is (id bigint, data string), making this example invalid.
Table was defined as local.db.table not local.db.target so updated this in MERGE INTO

What changed

Updated the example to use the correct data column:

MERGE INTO local.db.table t USING (SELECT * FROM updates) u ON t.id = u.id
WHEN MATCHED THEN UPDATE SET t.data = u.data
WHEN NOT MATCHED THEN INSERT *;

[huaxingao]

To make this truly runnable (and align with TestRoundTrip’s TODO about the getting-started doc not being runnable), should we change local.db.target → local.db.table and define updates before the MERGE?

[varun-lakhyani]

Thanks. Updated the Getting Started doc to use local.db.table instead of local.db.target and added explicit CREATE TABLE and INSERT INTO statements for the source and updates tables to make the example runnable.

[huaxingao]

shall we also update 3.4/4.0/4.1?

[varun-lakhyani]

Updated. Removed the TODO comments from all Spark versions (3.4, 3.5, 4.0, 4.1).

[huaxingao]

Thanks @varun-lakhyani for the PR!

[varun-lakhyani]

Thanks @varun-lakhyani for the PR!

Thanks for the review and merging this.
Much appreciated.