正则表达式还是split（''）？

Question

I am completely at a lost with how to appropriately construct a regular expression to do what I want with this file.

https://www.dropbox.com/s/9zadqzbvcg6ogtf/000218.txt?dl=0

AppearanceDate 29.08.2015
AppearanceTime 00:02:18
FrameCount 17
#   time  bright   x      y      alpha     delta   c_x    c_y    c_alpha  c_delta  use
01  18.175 ----  0.052  0.838   19.3755   21.947  -----  -----  --------  -------  no
02  18.215 ----  0.053  0.834   19.3682   21.985  -----  -----  --------  -------  no
03  18.255 ----  0.055  0.830   19.3608   22.024  -----  -----  --------  -------  no
04  18.295  5.1  0.057  0.826   19.3535   22.063  -----  -----   19.3541   22.070  yes
05  18.335  0.4  0.058  0.821   19.3462   22.101  -----  -----   19.3452   22.105  yes
06  18.375  0.3  0.060  0.815   19.3354   22.137  -----  -----   19.3365   22.140  yes
07  18.415  0.3  0.061  0.811   19.3281   22.172  -----  -----   19.3278   22.174  yes
08  18.455  0.2  0.063  0.806   19.3193   22.210  -----  -----   19.3192   22.208  yes
09  18.495  0.2  0.064  0.801   19.3110   22.236  -----  -----   19.3107   22.241  yes
10  18.535  0.2  0.066  0.795   19.3018   22.286  -----  -----   19.3023   22.274  yes
11  18.575  0.1  0.068  0.791   19.2935   22.312  -----  -----   19.2939   22.306  yes
12  18.615 ----  0.069  0.786   19.2861   22.335  -----  -----  --------  -------  no
13  18.655 -0.1  0.070  0.782   19.2788   22.359  -----  -----   19.2776   22.369  yes
14  18.695 -0.1  0.071  0.776   19.2686   22.391  -----  -----   19.2695   22.400  yes
15  18.735 ----  0.073  0.770   19.2583   22.424  -----  -----  --------  -------  no
16  18.775 ----  0.074  0.764   19.2480   22.456  -----  -----  --------  -------  no
17  18.815 ----  0.076  0.758   19.2383   22.488  -----  -----  --------  -------  no

I would like to match both the HH:MM:SS from AppearanceTime, and the SS.sss, from under the "Time" column.

Currently I can almost do it in two steps - firstly for AppearanceTime I can use:

r"(\\d{2}:\\d{2}:\\d{2})"

As far as I've got with the SS.sss values is:

r"(\\d{2}[.]\\d{3})"

but this matches part of the values in AppearanceDate, alpha, delta, c_alpha and c_delta also.

Finally, just in case it matters - I've been testing here: https://regex101.com/ with the global and multiline flags on.

If anyone could help me out with this it would be most appreciated. There seem to be a load of good resources to help with regex creation but I am getting absolutely nowhere with it!

Another idea I had was that I could probably use split(' ') quite effectively for the SS.sss but I wanted to ask whether anyone has an idea of which of regex or split is more efficient as this will be applied to many thousands of files like the one given above.

Thanks a lot!

Answer 1

You may use

(?:AppearanceTime\s+|^\d+\s+)(\d{2}:\d{2}:\d{2}|\d{2}\.\d{3})

See the regex demo (use the re.M flag with re.findall ).

Details :

• (?:AppearanceTime\\s+|^\\d+\\s+) - this matches 2 alternatives
  • AppearanceTime\\s+ - AppearanceTime strings and 1+ whitespaces ( \\s+ )
  • | - or
  • ^\\d+\\s+ - start of a line ( ^ ), 1+ digits ( \\d+ ) and 1+ whitespaces
• (\\d{2}:\\d{2}:\\d{2}|\\d{2}\\.\\d{3}) - matches and captures (the final output for re.findall ) either of the 2 alternatives:
  • \\d{2}:\\d{2}:\\d{2} - 3 : -separated 2-digit chunks
  • | - or
  • \\d{2}\\.\\d{3} - 2 digits, . , 3 digits substring

See the Python demo :

import re
rx = r"(?:AppearanceTime\s+|^\d+\s+)(\d{2}:\d{2}:\d{2}|\d{2}\.\d{3})"
s = <<YOUR STRING HERE>>
res = re.findall(rx, s, flags=re.MULTILINE)
print(res)

Answer 2

match = re.findall(r'^\d.+?(\d{2}[.]\d{3})', txt, flags=re.MULTILINE)
print(match)

out:

['18.175', '18.215', '18.255', '18.295', '18.335', '18.375', '18.415', '18.455', '18.495', '18.535', '18.575', '18.615', '18.655', '18.695', '18.735', '18.775', '18.815']

just use multiline mode, each line match first appear by ^\\d^\\d.+? .