name (DataStreamReader)

Assigns a name to the streaming source for checkpoint evolution. This allows streaming queries to evolve by enabling sources to be reordered or added without breaking checkpoint compatibility. When source evolution is enabled, all sources in a query must be named.

Syntax

name(source_name)

Parameters

Parameter	Type	Description
source_name	str	The name for this streaming source. Must contain only ASCII letters (a-z, A-Z), digits (0-9), and underscores (_).

Returns

DataStreamReader

Notes

Requires streaming source evolution to be enabled via the spark.sql.streaming.enableSourceEvolution configuration.

Examples

Name two sources and union them:

df1 = spark.readStream.format("rate").name("source1").load()
df2 = spark.readStream.format("rate").name("source2").load()
query = df1.union(df2).writeStream.format("console").start()

Valid and invalid source names:

# Valid names
spark.readStream.format("rate").name("mySource").load()
spark.readStream.format("rate").name("my_source_123").load()

# Invalid name — raises AnalysisException
spark.readStream.format("rate").name("my-source").load()