replace full stop by comma

Question

For the below values I have to replace the second and fourth "." by "," in the below file

input

1.351364711.103.7319660.2010-01-01 00:00:00
1.345529841.103.7372875.2010-01-01 00:00:49
1.342955629.103.7455272.2010-01-01 00:01:42
1.339694956.103.7520503.2010-01-01 00:02:28

desired intermediate output

    1.351364711,103.7319660,2010-01-01 00:00:00
    1.345529841,103.7372875,2010-01-01 00:00:49
    1.342955629,103.7455272,2010-01-01 00:01:42
    1.339694956,103.7520503,2010-01-01 00:02:28

I know awk gsub(/./,",") this replaces everything by comma.But I only need the columns to be seperated by "," . I also wanted to switch the third column in the first place after this.

desired final output

2010-01-01 00:00:00,1.351364711,103.7319660
2010-01-01 00:00:49,1.345529841,103.7372875
2010-01-01 00:01:42,1.342955629,103.7455272
2010-01-01 00:02:28,1.339694956,103.7520503

Answer 1

bash-4.1$ cat file
1.351364711.103.7319660.2010-01-01 00:00:00
1.345529841.103.7372875.2010-01-01 00:00:49
1.342955629.103.7455272.2010-01-01 00:01:42
1.339694956.103.7520503.2010-01-01 00:02:28

bash-4.1$ awk -F. '{print $NF,$1"."$2,$3"."$4}' OFS=, file
2010-01-01 00:00:00,1.351364711,103.7319660
2010-01-01 00:00:49,1.345529841,103.7372875
2010-01-01 00:01:42,1.342955629,103.7455272
2010-01-01 00:02:28,1.339694956,103.7520503

Answer 2

Using sed:

$ sed 's/\./,/4; s/\./,/2' file
1.351364711,103.7319660,2010-01-01 00:00:00
1.345529841,103.7372875,2010-01-01 00:00:49
1.342955629,103.7455272,2010-01-01 00:01:42
1.339694956,103.7520503,2010-01-01 00:02:28

This first replaces the 4th dot with a comma, then the 2nd.

We could have done the substitutions in the opposite order too, but since replacing the 2nd dot removes a dot from the line, the dot originally being 4th becomes the 3rd dot:

$ sed 's/\./,/2; s/\./,/3' file
1.351364711,103.7319660,2010-01-01 00:00:00
1.345529841,103.7372875,2010-01-01 00:00:49
1.342955629,103.7455272,2010-01-01 00:01:42
1.339694956,103.7520503,2010-01-01 00:02:28

The s/.../.../n syntax (where n is a digit) substitutes the n:th matching occurrence on a line.

Moving the last column to the front may be done with

s/^\(.*\),\([^,]*\)/\2,\1/

I.e., match stuff from the start of the line up to a comma, then the comma, then stuff that contains no more commas.

So the full command may be

$ sed 's/\./,/4; s/\./,/2; s/^\(.*\),\([^,]*\)/\2,\1/' file
2010-01-01 00:00:00,1.351364711,103.7319660
2010-01-01 00:00:49,1.345529841,103.7372875
2010-01-01 00:01:42,1.342955629,103.7455272
2010-01-01 00:02:28,1.339694956,103.7520503