AWK associative array, mapping

问题

Suppose I have two files:

file1 - map.txt

1, 178246
2, 289789
3, 384275
4, 869282

file2 - relation.txt

178246, 289789
384275, 178246
384275, 869282

Expected results are:

1, 2
3, 1
3, 4

But the results I got using the following code were:

awk 'FNR==NR{map[$2]=$1} {$1=map[$1];$2=map[$2];print $0}' map.txt relation.txt

  2,
  1,
  4,

It was confused when I swapped the columns in map.txt like this:

178246, 1
289789, 2
384275, 3
869282, 4

relation.txt doesn't change

The results became:

awk 'FNR==NR{map[$1]=$2} {$1=map[$1];$2=map[$2];print $0}' map.txt relation.txt

1,
3,
3,

It seems that something wrong near {$1=map[$1];$2=map[$2];print $0}

回答1:

Remove the leading space in both files in column 2.

And do yourself a favor and switch to something besides commas for FS. The Tab char is good because most input screens use tab to move to the next field, so it shouldn't be in your data. The | char is nice because it is visual and very unlikely to be in your input.

You could build a 'trap' to find records without the right number of fields like this:

awk -F"|" -v expectFldCnt=2 '{
   if (NF==expectFldCnt) { print ":everything OK" ; }
    else { print "ERR: " NF "!=" expectFldCnt  ":" $0 > "errorFile" }
    }' \
  map.txt relation.txt

IHTH

回答2:

awk  -F"[, ]" 'NR==FNR {m[$3]=$1;next};{print m[$1]",",m[$3]}' map.txt relations.txt

来源：https://stackoverflow.com/questions/21105473/awk-associative-array-mapping