cassandra-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sotirios Delimanolis <>
Subject Re: Atomic behavior and efficiency of a DELETE query with an IN clause
Date Wed, 10 Jun 2015 18:26:18 GMT
Will this "eventually they will all go through" behavior apply to the IN? How is this query
written to the commitlog?
Do you mean prepare a query likeDELETE FROM MastersOfTheUniverse WHERE mastersID = ?;and execute
it asynchronously 3000 times or add 3000 of these DELETE (bound) prepared statements to a
BATCH statement executed asynchronously?

     On Wednesday, June 10, 2015 9:51 AM, Jonathan Haddad <> wrote:

 Batches don't work like that.  It's possible for some to succeed, and later, the rest will. 
Atomic is the incorrect word to use, it's more like "eventually they will all go through".

Do not use IN(), use a whole bunch of prepared statements asynchronously.  
On Wed, Jun 10, 2015 at 9:26 AM Sotirios Delimanolis <> wrote:

When executing a DELETE statement with an IN clause, where the list contains partition keys,
what is the underlying behaviour with regards to atomicity?
DELETE FROM MastersOfTheUniverse WHERE mastersID IN ('Man-At-Arms', 'Teela');
Is it going to act like an atomic batch where if one fails, all fail? If that is the case,
is there any reason to use a BATCH statement with multiple single DELETE statement or should
we always prefer a DELETE with an IN clause? 
For example, given 3000 keys for rows I want to delete, should I issue a single DELETE query
and provide all the keys in the IN argument or should I add 3000 DELETE queries to a BATCH
Thank you,Sotirios

View raw message