Hello,

Any help Or idea?

Thanks


From: alioujigga@hotmail.com
To: user@cassandra.apache.org
Subject: INserting data in Cassandra
Date: Wed, 11 Apr 2012 09:42:52 +0000

Hello all,
We would like to adopt Cassandra solution for storing our biological data which are essentially microarray data.
These data, formatted in text tabulated files, are in the form:

               Sample 1             sample 2                         sample n

probe 1    value1                  value 2                            

probe 2         

probe n


In fact the probes can vary from one chip to another and you can have chips with more than one million probes, the samples by cons vary from one project to another, the values ​​are floats.
So we would like to represent our column families like they are formatted in files with the names of the probes (which are unique) as the key, and the names of the samples as column names and float values as column values :)
To insert data, I just created such for example a keyspace testKS and a column family testCF but just by defining the key (because the column names vary), then a file jason for insertion:

{"probe1" :{

                "sample 1":value1(float),

                "sample 2":value2(float),

                …,

                "sample n":value n(float)

},

"probe2" :{

                "sample 1":value1(float),

                "sample 2":value2(float),

                …,

                "sample n":value n(float)

}

}


 
And I used the tool json2sstable, but that does not work, I always have an error:

java.lang.RuntimeException: Can't write Super columns to the Standard Column Family.

So I have two questions:
1) What I did wrong, must I define the complete structure of my column family before using json2sstable or is it the structure of my json file which is not good?
2) Otherwise what would be the best way to proceed for insertion based on the data I dispose?

We use Cassandra 1.0.8. Any help would be welcome.

Thanks.