Converting Large scale Daily Precipitation Values to Monthly Values

Hello folks! I have to convert very huge scale daily rainfall values to monthly values. The data comprises of more than 2000 stations and for 20 years of daily values. Processing this much extent of data is not possible in Excel. Can someone suggest how to deal with it in MATLAB?

[VINAYAK LUHA]

Hi Faisal ,

Could you please provide more info on defination of your dataset ,attributes,whether it has NULL/NAN values or not etc.

[Faisal Baig]

its the rainfall data for almost 2400 stations and comprises of almost 20 years of daily values. It might have some NaN values as well for places and days when there was no rainfall recorded. I want to convert those daily values to monthly accumulative values which is very difficult in excel considering the hugeness of the data.

[dpb]

Sure. Piece o' cake. Use readtimetable to bring in the data and then standardize the missing values with standardizeMissing and compute to heart's content with retime

And, that's not all that much data in today's terms of GB storage;

>> 20*365*2000*8/1024/1024
ans =
  111.3892
>> 

is ~112 MB. For precipitation data, you can cut storage memory in half -- single precision will be more than enough precision so the datetime array for the timetable is only ~60 KB. I've not deeply investigated the overhead of the timetable itself, but I don't think you should have a problem with a reasonable machine to use.

See <function-of-rainfall-seasonality-index-walsh-and-lawler-1981#comment> that illustrated how to get started with some precipitation data for another purpose; that's about the end goal for your above problem.

[Faisal Baig]

thanx alot. That was really helpful