At what isolation level are Snowflake streams?

The correct answer is B. Repeatable read. Snowflake streams operate at Repeatable Read isolation because once a transaction begins consuming a stream, all reads within that transaction see the same consistent set of change records - the stream's offset is fixed at transaction start and won't shift if other transactions…

Data Movement

Question

Options

ASnapshot
BRepeatable read
CRead committed
DRead uncommitted

How the community answered

(37 responses)

A
5% (2)
B
92% (34)
C
3% (1)

Explanation

Snowflake streams operate at Repeatable Read isolation because once a transaction begins consuming a stream, all reads within that transaction see the same consistent set of change records - the stream's offset is fixed at transaction start and won't shift if other transactions commit in the meantime. Snapshot isolation (A) is a related but distinct concept (used in databases like SQL Server/Postgres); Snowflake's documentation specifically uses the term "repeatable read" for streams, not snapshot. Read Committed (C) would allow each successive read within a transaction to see newly committed data, which would break stream consistency - consumers would get a moving target. Read Uncommitted (D) would expose dirty (uncommitted) writes, which is fundamentally incompatible with stream semantics where only committed DML changes appear.

Memory tip: Think of "REPEAT" - stream consumers can re-read the same change records within a transaction and get the same results every time, because the view is locked at transaction start.

Topics

#Snowflake Streams#Isolation Level#Repeatable Read#Change Data Capture

Community Discussion

No community discussion yet for this question.

Full DEA-C02 Practice