Home / ubuntu / Questions / 1109590
Accepted
deepblue_86
Asked: 2019-01-15 02:14:34 +0800 CST

extracting numbers whose lengths are fixed from text file

  • 772

I have a text file. This file includes characters and numbers as follows:

ANKR00TUR_R_20183240000_01D_30S_MO.rnx:  2018    11    20    00    00    0.0000000     GPS         TIME OF FIRST OBS
brmu3350.14o:  2014    12     1     0     0    0.0000000     GPS         TIME OF FIRST OBS
KNY12040.14o:  2014     7    23     0     0    0.0000000     GPS         TIME OF FIRST OBS
rinex_quantity:grep "TIME OF FIRST OBS" * > time_of_first_epochs

I need to extract only 4 digits numbers and store them into another file as follows:

2018
2014
2014

I applied the following code but it extracts all 4 digit numbers:

grep  -Po "\d{4}" data

2018
3240
2018
0000
3350
2014
0000
1204
2014
0000
text-processing

2 Answers

  1. Best Answer

    Your grep command was almost correct, you just have to anchor the pattern to match only if there is a word boundary before or after it.

    Word boundaries are zero-length patterns that match between a word-character (letters, digits, underscore) and a non-word charater (e.g. spaces, other punctuation, line end, and everything else).

    In grep, you can either do this by surrounding your pattern with \b, or by using the -w switch to enable word matching:

    $ grep -Po '\b\d{4}\b' data
2018
2014
2014

$ grep -Pow '\d{4}' data
2018
2014
2014
    • 7


  2. with miller (http://johnkerl.org/miller/doc) is 

    mlr --implicit-csv-header --pprint  cut -f 2 then label year input

    As output you will have

    year
2014
2014

    Mi input is

    brmu3350.14o:  2014    12     1     0     0    0.0000000     GPS         TIME OF FIRST OBS
KNY12040.14o:  2014     7    23     0     0    0.0000000     GPS         TIME OF FIRST OBS

    I have simply extracted the second column with cut

    • 0