Cassandra Primary Key allows duplicates

Question

I tried inserting two record with all same values for columns in Primary Key in a Cassandra table and insert was successful. I am new to Cassandra and thought Primary Keys in Cassandra would not allow duplicate insertion. Is that incorrect? Following are the columns in my Primary key

PRIMARY KEY ((customer_id, source_id ), status_code, create_timestamp, modified_timestamp)

Following is how I am inserting

insert into testkeyspace.customers 
(customer_id, source_id, status_code,
 create_timestamp, modified_timestamp) 
value ('123e4567-e89b-12d3-a456-426655440000', 
1122334455, 0, toTimestamp(now()), toTimestamp(now()));

Robbie · Answer 1 · 2018-03-21T20:44:28.380

3

You're using toTimestamp(now()) for 2 of the clustering columns in your primary key, so its unlikely that the 2 inserted records have exactly the same values for those columns, i.e. the two records are not duplicates.

Take a look at this answer to a similar question

edited Mar 21 '18 at 20:44

answered Mar 21 '18 at 20:39

Robbie

18,750
4
41
45

1

Beyond this, cqlsh may report them as having the same timestamp, but some versions of cqlsh don't show milliseconds. – Jeff Jirsa Mar 21 '18 at 21:51
it's really depends on your configuration of `time_format` in `~/.cassandra/cqlshrc` – Alex Ott Mar 22 '18 at 08:27

score 3 · Accepted Answer · answered Mar 21 '18 at 20:46

An important thing to note is the C* does not specifically disallow having duplicate values for a primary key. Having a duplicate value will overwrite the previous value (based on writetime), there will be no client messages to indicate this is happening. You will only notice on subsequent reads.

Cassandra Primary Key allows duplicates

2 Answers2