Suppress numpy warnings?

[durack1]

I've just started seeing these warnings:

import cdms2 as cdm
import MV2 as MV
# Set nc classic as outputs
cdm.setNetcdfShuffleFlag(0)
cdm.setNetcdfDeflateFlag(0)
cdm.setNetcdfDeflateLevelFlag(0)

os.chdir('/work/durack1/Shared/131118_PaperPlots_SaltyVariability/151001_NOEMP02_IPSL-CM6')
domains = ['glo','atl','pac','ind']
for i in domains:
    var = 'zosal' + i
    infile = ''.join(['NOEMP02_1850_1899_1Y_',var,'.nc'])
    print '** Processing: ',infile
    f_h = cdm.open(infile)
    var1850to1899 = f_h(var) ; f_h.close()
    infile = replace(infile,'1850_1899','1900_1949')
    f_h = cdm.open(infile)
    var1900to1949 = f_h(var) ; f_h.close()
    vars()[i] = MV.concatenate([var1850to1899,var1900to1949])
    del(var1850to1899,var1900to1949,var,infile) ; gc.collect()
del(i,domains) ; gc.collect()

>>> glo_1850to1899.shape
(50, 31, 149, 1)
>>> glo_1900to1949.shape
(50, 31, 149, 1)
>>> glo = MV.concatenate([glo_1850to1899,glo_1900to1949])
/usr/local/uvcdat/2.4.0rc1/lib/python2.7/site-packages/cdms2/axis.py:2266: FutureWarning: comparison to `None` will result in an elementwise object comparison in the future.
  if None in boundsArray:

It will likely disappear in planned numpy/cdms updates #885 (and has already been noted before #576), but wondering whether this should be cleaned up before 2.4.0?

[dnadeau4]

@durack1 I get no warnings!! Can you try this code?

import cdms2
import numpy
aa=cdms2.MV.array(numpy.ma.ones([50,31,149,1]))
bb=cdms2.MV.array(numpy.ma.ones([50,31,149,1]))
cc=cdms2.MV.array(numpy.ma.ones([50,31,149,1]))
aa.shape
(50, 31, 149, 1)
bb.shape
(50, 31, 149, 1)
cc.shape
(50, 31, 149, 1)
dd=cdms2.MV.concatenate([aa,bb,cc])
dd.shape
(150, 31, 149, 1)

[durack1]

@dnadeau4 it seems it has something to do with the input data (and the coordinate bounds), I'm guessing as I got similar results to you below:

In [1]: import cdms2
In [2]: import numpy
In [3]: aa=cdms2.MV.array(numpy.ma.ones([50,31,149,1]))
In [4]: bb=cdms2.MV.array(numpy.ma.ones([50,31,149,1]))
In [5]: cc=cdms2.MV.array(numpy.ma.ones([50,31,149,1]))
In [6]: aa.shape
Out[6]: (50, 31, 149, 1)
In [7]: bb.shape
Out[7]: (50, 31, 149, 1)
In [8]: cc.shape
Out[8]: (50, 31, 149, 1)
In [9]: dd=cdms2.MV.concatenate([aa,bb,cc])
In [10]: import MV2 as MV
In [11]: ee = MV.concatenate([aa,bb,cc])

For completeness I updated my code snippet above..

[dnadeau4]

Fixed using new numpy and with new releases.