awk command for filtering

Question

[nmaden]

I have the following input data in .tsv format (a piece of example from a file that contains >100k rows/lines).

Chr:Pos ID  REF ALT Q   AD  DP  GQ  GT
1:1246257       GGCT    -   62.73   31,5    36  99  0/1
1:1267678       TGCA    -   0   ,           ./.
1:17085502  rs2329006   T   C   1171.77 20,52   72  99  0/1
1:17085564  rs3851921   A   G   446.77  21,19   40  99  0/1
1:17085590  rs371449598 GCGCTG  -   88.73   37,12   49  99  0/1
1:17085594      T   C   0   ,           ./.
1:119964933 rs75429891  T   C   373.77  26,19   45  99  0/1
1:120057158 rs6203  C   T   1351.77 18,21   39  99  0/1
1:111783982 rs13721 C   A   782.77  0,32    62  99  1/.
1:111783982 rs13721 C   T   782.77  0,30    62  99  ./1
1:152195729 rs34061715  T   -   576.73  0,17    21  38  ./1
2:73613032  rs193922695 -   GGA 786.73  0,8 21  99  ./1
2:73613032  rs61156725  GGAGGA  -   615.73  0,13    21  99  1/.
2:73613032  rs61156725  GGAGGA  -   615.73  0,,13   21  99  2/.
2:241696841 rs71779512  TCC -   944.73  0,8 19  99  ./1
3:42251578      -   GGA 1128.73 0,8 28  99  ./1
3:42251578  rs10530663  -   GGA 1128.73 0,2,8   28  99  ./2
1:981931    rs2465128   A   G   1100.77 10,37   47  99  0/1
1:982994    rs10267 T   C   2661.77 54,28   82  99  0/1
1:984302    rs9442391   T   C   1054.77 50,28   78  99  0/1
1:1007203   rs4633229   A   G   372.77  9,7 16  99  0/1
1:1007432   rs4333796   G   A   626.77  22,16   38  99  0/1
1:1246257       GGC -   62.73   31,5    36  99  0/1
1:1290276   rs75904949  C   G   231.77  10,12   22  99  0/1

I want to filter out GT records with ./. , 1/., ./1, 2/., ./2 and print the whole table.

For this I used the awk command as below:

awk '($9!="./." && $9!="1/." && $9!="./1"  && $9!="2/." && $9!="./2") {print $0}'  input.tsv > output.tsv

The command runs successfully but the output file still has the records with the above characters.

Any help is appreciated.

Thanks

[Pierre Lindenbaum]

but the output file still has the records with the above characters.

show us such lines.

[Pierre Lindenbaum]

also test awk -F '\t' '($9!="./." etc ...

[lieven.sterck]

I don't think this will work as intended.

When you use the 'negative' match ( != ) you more likely need to use the or operator ( || ) and not the and operator ( && ). Can you give that a try ?

EDIT: nevermind the above, just listen to Pierre Lindenbaum :)

[Pierre Lindenbaum]

it should work

echo -e '1.1\n./2' |  awk '($1!="./." && $1!="1/." && $1!="./1"  && $1!="2/." && $1!="./2")'
1.1

[lieven.sterck]

yeah, you're right. Let's see where OP takes this

(and I know better than to debate with you on awk topics, but I'm just in one of those moods ;) )

[nmaden]

Hi, the code can remove 1/., ./1, 2/. and ./2 but not ./. Tried varying it but no avail.

Thanks!