Shell Programming and Scripting

Hello,

I am looking for a specific situation in a text file. The conditions are,

Code:

>  <CompoundName>
InChI=1S/C5H12NO2/c1-5(2)4-8-6(3)7/h5H,4H2,1-3H3/q+1

I am looking for cases where the line "> <CompoundName>" is followed by a line that contains the string "InChI=" without regard to what follows InChI=.

I tried in awk,

Code:

$ cat ./filename | awk 'F = 1 { if($0 ~ "InChI=") { print $0; F = 0 } } $0 ~ "<CompoundName>" { F = 1 } '

This is intended to mean that when $0 contains "<CompoundName>", F is set to 1. When F = 1 (the line following "<CompoundName>"), if the line contains "InChI=", the line is printed and F is reset.

This does not work and I get a print of every line that contains "InChI=".

What did I do wrong here? I would guess that grep is the better tool for this but I don't know grep well.

Thanks,

LMHmedchem

Code:

F == 1

Try also

Code:

awk '/<CompoundName>/ {getline TMP; if (TMP ~ /InChI=/) print TMP}' file

or even

Code:

awk '/<CompoundName>/ {getline; if (/InChI=/) print}' file

You can also try with sed.

Code:

sed -n '/<CompoundName>/!d;n;/InChI=/p' infile

When changing to,

cat inputfile | awk 'F == 1 { if($0 ~ "InChI=") { print $0; F = 0 } } $0 ~ "<CompoundName>" { F = 1 } '

I get the same output running this as I do for F = 1. Do I have the braces wrong or something?

This following solutions works as far as tested,

sed -n '/<CompoundName>/!d;n;/InChI=/p' infile

awk '/<CompoundName>/ {getline TMP; if (TMP ~ /InChI=/) print TMP}' file

awk '/<CompoundName>/ {getline; if (/InChI=/) print}' file

Thanks for the help. I am surprised to not see a grep solution. I always think of grep as a way to find and print text.

LMHmedchem

Might be due to spurious F , mayhap set by a CompoundName way before. You should reset F after each line processed, like

Code:

awk 'F == 1 && /InChI=/; { F = 0 } $0 ~ "<CompoundName>" { F = 1 } ' file
InChI=1S/C5H12NO2/c1-5(2)4-8-6(3)7/h5H,4H2,1-3H3/q+1

grep prints lines FOUND (plus context, that is); it CAN'T print the following line only... If your grep allows for the -A option, try

Code:

grep -A1 CompoundName file | grep InChI

(implying the InChI is not contained in the CompoundName lines)