Extracting data from filenames in Gnuplot

Question

Is there a way for Gnuplot to read and recognize structured strings? Specifically, I have a few hundred files, all containing measurement data, with measurement conditions defined in the filename.

My files look something like "100d5mK2d0T.txt", which would mean that this data was acquired at 100.5mK temperature and 2.0T magnetic field.

Any chance I could extract the temperature and field strength data from the name, and use them as labels in the plot?

Thanks in advance.

How are you running your gnuplot script--interactively or calling a canned script repeatedly? Also, what OS/gnuplot version are you using? — andyras
– andyras, Commented Apr 4, 2014 at 4:08
I would suggest to use external parser. Gnuplot has the c-like sprintf function but not the scanf. Nevertheless some string manipulation is possible. As example name="100d5mK2d0T.txt"; temp = int(name[0:3])+int(name[5:6])/10.0; Bfield = int(name[8:9])+int(name[10:11])/10.0 will do what you want, but this implies the formatting is strictly fixed... It would'n work for strings like 99d5mK2d05T.txt — bibi
– bibi, Commented Apr 4, 2014 at 6:45

Christoph · Accepted Answer · 2014-04-04 11:50:34Z

2

With gnuplot's internal string processing you could come up with a solution (using substr and strstrt), but thats quite verbose.

Its better to use an external tool for the string processing, like perl:

filename = '100d5mK2d0T.txt'
str = system('echo "'.filename. '" | perl -pe ''s/(\d+)d(\d+)mK(\d+)d(\d+)T.txt/\1.\2 \3.\4/'' ')
temperature = word(str, 1)
magnetic_field = word(str, 2)

set label at graph 0.1,0.9 "Temperature: ".temperature." mK"
set label at graph 0.1,0.8 "Magnetic field: ".magnetic_field." T"

edited Apr 4, 2014 at 11:50

answered Apr 4, 2014 at 7:20

Christoph

48.7k8 gold badges94 silver badges198 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Zvonimir Domazet Over a year ago

Thank you, this is exactly what I needed! One more detail, if you don't mind. Besides the useful information, filenames also carry some junk at the end. Something like "49d0mKnow_50d0mKset_3d00T_D03_1_240_001.dat". So, I would like to extract the second temperature and the magnetic field, while ignoring everything else. I managed to do so by modifying your code a bit: perl -pe ''s/(\d+)d(\d+)mKnow_(\d+)d(\d+)mKset_(\d+)d(\d+)T(\s)*/\1.\2 \3.\4 \5.\6 \7/'' I'm not really familiar with Perl syntax, so I just wanted to ask if this is okay, or is there a better way to do it?

Christoph Over a year ago

I'm also not very familier with perl, just with regular expressions. As long as it works, its fine :) Depending on the different formats which your file names can have, it may be easier to use one perl call for each variable instead of extracting all together and then splitting them with word inside gnuplot.

Collectives™ on Stack Overflow

Extracting data from filenames in Gnuplot

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related