cube.load() is taking very long time. How can I make it faster?

[Penelope Maher]

Hi all,

I am new to iris and trying to load quite a large amount of UM data in .pp file format. My goal is up load 28 years of monthly data but for only a subset of stashcodes of interest (approx 15 variables) and output as .nc (ideally into one file but no problem if I need to join them later).

Each input file contains one month of monthly data and there are on the order of 200 variables in each file. My code takes about 2 minutes to run 4 months of data. So the projected run time is about 3 hours. Is there a faster way to do this? I could loop over years or similar, but that won't bring down the run time (just mitigate problems if the code crashes).  The code I am using to load the cube is at the end of the email.

Thank you,

Penelope

stash = ['03i217',  '03i234']

for var  in stash:

    stash_list.append('m01s{0}'.format(var))

stash_num = iris.AttributeConstraint(STASH=lambda x: x in stash_list)

cubes = iris.load(file_list, stash_num) #file_list is a list of strings for the file names