georinex
georinex copied to clipboard
geospace-code
Read .o files with extreme long time
I am trying to read a .o file with gr.read(), but it cost me very long time >10m, can any one helps me?
ista3090.18o.gz Sorry, I upload a wrong file. this one is ok. md5sum:6dca94165d45a455c0bb05b1b564ed68
Thanks @ZephyrShannon this is an issue we intend to resolve in the future. OBS2 has been massively sped up to near-C speed, but OBS3 still needs that updated algorithm. Stay tuned...
@ZephyrShannon we recently added the ability to read data from an OBS2 or OBS3 at every interval= seconds. For example if the big file has measurements once a second, you can read every 15 seconds or whatever suits your needs, giving a proportion speedup in reading.
The real solution of preallocating Numpy arrays has not yet been implemented for OBS3, but was several months ago for OBS2. I'm sorry for this delay, I just haven't had time to do this yet.
Dear scivision, regarding this update, I have found one rinex file that leaves an error of the type: IndexError(f'may be "fast" mode bug, try fast=False or "-strict" command-line option {e}') please, find the rinex file at ftp://gssc.esa.int/gnss/data/daily/2013/360/usno3600.13d.Z
Otherwise, if I remove the fast option, by saying fast=False in the function rinexobs2 when it calls the rinexsystem2, the rinex reading becomes correct again.
Thanks @fabprol for providing the link to the example data, I plan to look at this soon.
Hey scivision, indeed, the rinex I sent before is working normally, the one that is not really working is this one: https://drive.google.com/open?id=199UHtvZVkhXG6iMl07bNI4zeMnaR1r5D
which is the same, but converted to RINEX by crx2rnx and edited by Windows/TEQC using the following command: teqc.exe -S -O.obs C1P1P2L1L2 usno3600.13o > usno3600Edited.13o