incubator-cassandra-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ertio Lew <ertio...@gmail.com>
Subject Merging the rows of two column families(with similar attributes) into one ??
Date Sat, 05 Feb 2011 04:27:40 GMT
I read somewhere that more no of column families is not a good idea as
it consumes more memory and more compactions to occur & thus I am
trying to reduce the no. of column families by adding the rows of
other Column families(with similar attributes) as separate rows into
one.

I have two kinds of data for two separate features on my application.
If I store them in two different column families then both of them
will have similar attributes like same comparator type & sorting
needs. Thus I can also merge both of them in one column family, just
by adding the rows of another to this one(increasing the no of rows).
However some rows of 1st kind of data are very frequently used and
rows of 2nd data are less freq. used. But I dont think this will be a
problem as I am not merging two rows into one, but just adding them as
separate rows in the column family.
1st kind of data has wider rows and 2nd kind of data has very less wide rows.

But the caching requirements may be different as they cater to two
different features.(but I think it is even advantageous since
resources are free to be utilized by any data that's more frequently
used)


Is it recommended to merge these two column families into one ?? Thoughts ?

--

Ertio

Mime
View raw message