Awk insert new column in output file

Shell Programming and Scripting
1

Dear team,

Here is my input file below. I need awk script which insert new column $8 with same value as $7 but in case $7=blank then same value as previous value .

$cat sample.txt:

ACT     1J                      AD-300  TIME 200322 1554  HLRBC03
ORDERED

ACT     1J                      AD-300  TIME 200322 1554  HLRBC08
ORDERED

ACT     1J                      AD-300  TIME 200322 1554  HLRBC12
ORDERED

ACT     1J                      AD-300  TIME 200322 1554  HLRBC18
ORDERED

ACT     1J                      AD-300  TIME 200322 1554  HLRBC04
ORDERED

ACT     1J                      AD-300  TIME 200322 1554  HLRBC07
ORDERED

ACT     1J                      AD-300  TIME 200322 1554  HLRBC15
ORDERED

ACT     1J                      AD-300  TIME 200322 1554  HLRBC19
ORDERED

Like below

ACT     1J                      AD-300  TIME 200322 1554  HLRBC03 HLRBC03
ORDERED                                                                                                  HLRBC03
                                                                                                                    HLRBC03

I tried this

awk '{a[++i]=$0;}END{for (j=1;j<NR;j++) {if (a[j]~"HLRBC") {for (y=1;y<NR;y++) $7==$8 print $8}}}' sample.txt

Kindly need support.. really appreciated

2

the requirement and the provided sample input/output are a bit fuzzy and somewhat badly formatted.
See how far this will help you:

awk 'NF>1 {$(NF)=$NF OFS $NF}1' myFile
3

Hi

sed -r 's/\s+\w+$/&&/' myFile

Good when there is an alternative

1 Like
4

That works with many awk versions.
The following works with all awk versions:

awk '{ if (NF>1) print $0, $7; else print }' myFile

But nobody here yet understood the requirement "in case $7=blank then same value as previous value".

1 Like
5

@MadeInGermany, do you know what incarnation/version of awk does not handle what was proposed initially?

I cannot recall where this would fail....

That's correct. I couldn't understand what it meant and couldn't correlate it to the sample input file(s).

1 Like
6

Across the seventh field

sed -r 's/(\s*\S+){7}/&\1/' myFile
7

produces a funny ORDEREDD with GNU awk 4.0

Perhaps the O/P wants a repetition of the new column when there is no $7? Then consider the following:

awk '(NF==7) { newcol=$7 } { print $0, newcol }' MyFile

Or a bit finer formatted:

awk '(NF==7) { newcol=$7; newpos=length+16 } { printf "%s%*s\n", $0, newpos-length, newcol }' MyFile

@vgersh99, maybe your code works with all awk versions (besides the Solaris oawk of course), maybe there is only a difference what NF is after the forced reformatting.
But generally I think the forced reformatting is often not the best idea.

8 
awk '$7 {print $0,$7} !$7' myFile
9

Thanks a lot .. Really appreciate if you explain it as well. as i need slightly variation

here instead $7 need to put condition that newcolumn should be printed if $7 start with ^HLRBC ...

like in below example instead of RESTART value should be "HLRBC03"

ACT     1J                      AD-300  TIME 200322 1554  HLRBC03      HLRBC03
ORDERED                                                                                         HLRBC03
1 RELOAD        COMMAND ORDERED CP CLUSTER RESTART  RESTART
ACT     1J                      AD-300  TIME 200322 1554  HLRBC08     HLRBC08
ORDERED                                                                                        HLRBC08
10

Replace the condition (NF==7) (number of fields == 7) with ($7~/^HLRB/)
Also it is smarter to set the field width for the first %s (that is the $0).

awk '( $7 ~ /^HLRB/ ) { newcol=$7; newpos=length } { printf "%-*s  %s\n", newpos, $0, newcol }'

%-s left adjusted
%*s the * is the field width given by an extra argument
%-*s left adjusted and field width by argument

2 Likes
11

That's great..

Thanks a lot for your support.

12

dear MadeIn,

Can we have another new column which has count of number of times same value in column 8

like below

ACT     HR1J                      AD-300  TIME 200322 1554  HLRBC01  HLRBC01 3
ORDERED                                                     HLRBC01  HLRBC01 3
                                                            HLRBC01  HLRBC01 3
ACT     HR1J                      AD-300  TIME 200322 1554  HLRBC16  HLRBC16 2
ORDERED                                                     HLRBC16  HLRBC16 2
13

Dear Madelin,

Here i tried and its worked

awk '( $8 ~ /^HLRBC/ ) { newcol=T[$8]++; newpos=length } { printf "%*-s  %s\n", newpos, $0, newcol }'