ratal
commented
5 years ago Hi, When reading whole file content, mdfreader uses efficient numpy methods to read the channels. When a channel list is specified, these numpy methods are not performing anymore, consuming too much memory and computation for only one channel for instance --> so I developped a cython module dataRead that is doing this one by one channel reading. However, this cython module has to be properly compiled for you platform --> not done properly done in your environment, the warning you get. I see you use python3.6 and win10, I think there is wheel available from pip ? Cython could be an issue for some users therefore I implemented a less computationnaly performing fallback in pure python (but still saving memory) --> you can still read your channel.
proedig
danielhrisca
Python version
'3.6.5 |Anaconda, Inc.| (default, Mar 29 2018, 13:32:41) [MSC v.1900 64 bit (AMD64)]'
Platform information
'Windows-10-10.0.14393-SP0'
Numpy version
'1.14.3'
mdfreader version
mdfreader 2.7.8
Description
Hi! I have problem with reading in mdffiles using the optional argument "channellist".
When I try
data = mdfreader.mdf(file)there is no problem. All channels are read in. When I try the same file with
data = mdfreader.mdf(file,channelList=['n'])I get the following output:
_C:\ProgramData\Anaconda3\lib\site-packages\mdfreader\mdf3reader.py:552: UserWarning: Unexpected error: (<class 'ModuleNotFoundError'>, ModuleNotFoundError("No module named 'dataRead'",), <traceback object at 0x00000258E3768EC8>) warn('Unexpected error: {}'.format(excinfo())) C:\ProgramData\Anaconda3\lib\site-packages\mdfreader\mdf3reader.py:553: UserWarning: dataRead crashed, back to python data reading warn('dataRead crashed, back to python data reading')
However, there is reasonable output:
_In [29]: data Out[29]: {'time_2': {'unit': 's', 'description': '', 'master': 'time_2', 'masterType': 1, 'data': array([5.60005003e-03, ..., 4.75285915e+03])}, 'n': {'unit': 'rpm', 'description': 'Engine speed', 'master': 'time2', 'masterType': 1, 'data': array([0, ..., 0], dtype=uint16)}}
So this means for me there is no problem with the variable 'n'. Is it a problem with cython? I have installed cython version 0.28.5.
data.MDFVersionNumber = 300
Best wishes
Philip package_list.txt