UNIX for Advanced & Expert Users

Hi,

I am looking for an opensource alternate to replace syncsort. Can you please suggest ?

Believe it or not, we replaced syncsort with plain the UNIX sort command, and got comparable results. syncsort was still about 10% faster, which did not justify the cost differential, USD $60K vs zero.

Hi Jim,

Thanks for your response. Apart from sorting we require to perform operations like "merge - combining as many as sequencial sets into one data set under same sequence" and "copy - to reproduce a file, bypassing the sorting process entirely"

So can we achieve these also ?

Thanks,
Ambika

You can usually achieve the "copy" procedure with "cp". Or if you're sticking a bunch together, "cat".

UNIX sort also has a 'merge' option, check the manual page to see if it can be used exactly in the manner you require.

The cp and cat functions Corona mentions are more efficient than the syncsort version.
Merge is simple as well: merge filea and fileb giving filec with the merge key being two fields pos 8-30 and pos 43-49:

Code:

sort -k 8,40 -k 43,49   filea fileb > filec

Ok. Thank you. I will try out these.

hello AmbikaValagonda,

I'm from Syncsort, and saw your thread and the answer you got.
I'm not going to emphasize how good our product is bla bla bla.
Nevertheless, let me just add a few words.

Syncsort SORT is faster than the other sorts when volumes are BIG, i.e. when the source data don't fit into memory. Then, we fear no comparisons.
(and i don't speak about 10%, but 30% to 60%)

Be also aware that each version of Syncsort is faster than the previous ones. If you're using Syncsort 3.X, you may want to give a try to v7.1 before going further.

Syncsort has implemented new features like workspace compression which could accelerate your processes and save additional disk space.

And last, we now have a graphical interface which could reduce your dev and maintenance time, helping you to make more savings than by implementing and maintaining unix shell scripts.

Syncsort solves problems, especially performance problems when data volumes are BIG.

Of course, young geeks prefer scripting. Being able to handle pattern language makes them feel smart.

Best regards,
Steven