CDO

DS Bisht wrote:

I have downloaded GLDAS-1 (3-hourly Noah 0.25 data) for the year 2013. As the data is at 3 hour time step I got 365*8=2920 files. When I used merge command in CDO to merge all these files into a single time series it failed citing "Too many open files" in reason. To overcome this problem I tried to merge all these files by segregating them into three sets and subsequently merging into one single time series, but the final time series is giving processed variables only on 2661 time steps which should have been on 2920 time steps.

Can any one tell me how to resolve this error ?

Quite often while executing Merge command I get warnings like

"cdo merge (Warning): Input stream 1 has 1 timestep. Stream 266 has more timesteps, skipped!"
What does it mean ?

The CDO operator merge is not intended to merge time series with the same variable to one single time series. See https://code.zmaw.de/projects/cdo/embedded/1.6.4/cdo.html#x1-820002.2.5 for more information.
You have to use the operator cat to concatenate time series.

Thanks for the reply.

I tried with operator cat too but it failed

cdo cat $(ls *.grb) TS.grb
cdo cat: 8%
cdo cat (Abort): Input streams have different number of variables per timestep!

This problem persists only for few months of the year (Feb, June, July and Sep), I am able to use any of the operator (merge or cat) successfully for other months. Is the problem with the downloaded data ?

The CDO operator cat requires that all input files have the same number of variables. It seems that some of your files do not fulfill this requirement.

I downloaded this data from http://disc.sci.gsfc.nasa.gov/hydrology/data-holdings
If there is any problem in the data then how to resolve this ? There must be some solution, at least some way to know for which date data is missing when extracting the time series for our analysis, so that we can assume it as missing data point.
Here I am attaching the downloaded files for the month of Feb, please have a look.

DS Bisht wrote:

I downloaded this data from http://disc.sci.gsfc.nasa.gov/hydrology/data-holdings
If there is any problem in the data then how to resolve this ? There must be some solution, at least some way to know for which date data is missing when extracting the time series for our analysis, so that we can assume it as missing data point.
Here I am attaching the downloaded files for the month of Feb, please have a look.

I have herewith attached the data for the month of Jan too, you can see we can use Merge and Cat operator for this month unlike Feb.

The CDO option -v prints the processed files:

cdo -v cat Feb/* result