sorting csv file with subfield

Question

I have 1 csv file space delimited,

Jan_high 32 123
Febr_low 19 139
March_high 12 63
Jan_low 36 18
Febr_high 87 99
March_low 83 77

how do I sort it based on _high, then _low string 1st column? I cant use sort -k1.5 since the string high or low doesn't start at fixed character location. Does anyone have idea ?

Answer 1

Using -t '_' -k2 with sort would make sort consider the input as delimited on the _ character instead of on the spaces, and then sort on the second field (the one starting with either high or low):

$ sort -t '_' -k2 file
March_high 12 63
Jan_high 32 123
Febr_high 87 99
Febr_low 19 139
Jan_low 36 18
March_low 83 77

Taking the first line of input as example:

Jan_high 32 123
^^^ ^^^^^^^^^^^
f1  field2