UNIX for Dummies Questions & Answers

I am relatively new to UNIX and am having some really weird results with redirecting output. I will start at the beginning.

I have a file with consists of a list of a chemical name followed by a tab and then a string of characters known as a SMILE string in this format

Code:

PS_1    C=C
PS_2    C-C-N
PS_3    C-C=C

the list is 22 lines long.

I am using csplit to split the file (the file is called 22hits) into multple individual files using this command:

Code:

csplit -skf seq 22hits '/PS_/' {20}

BTW, I am confused about why I have to use the number 20 instead of 22 if I want 22 separate files. Regardless, that is something I figured out by trial and error. This is not the major problem, however.

The resulting files have the name seq00, seq01, seq 02 etc. and contain and individual line of data including the name of the chemical and the SMILE string.

Now, to rename the files to the chemical names, I am using the following shell script:

Code:

#!/bin/sh
#Rename split chemical files
file="$1"
    fastaname=`cat $1 | perl -pi -e 's/\t/\n/g' | grep 'PS_'`
    echo $fastaname
    mv $1 $fastaname

This script changes the tab between the name of the chemical and the SMILE string to a newline, retrieves the name alone and then renames the file with that name.

Additionally, I am trying to remove the name of the compound in the file so that I have a file named PS_1 for example, and the only information it contains is the SMILE string, not the name of the compound, followed by a tab, followed by the SMILE string.

To accomplish this, I am using the following script:

Code:

#!/bin/sh
cat $1 | perl -pi -e 's/\t/\n/g' | grep -v 'PS_' > $1

Again, I am replacing the tabs with newlines, returning the line that does not have 'PS_', which should be the smile string, and then directing that output to the file. The problem is....for some files, i get the desired results, and for others, the file comes back with all data removed. The bigger problem is, I can repeatedly run this series of scripts on the same file (22hits) and get different results every time. Sometimes, half have the desired result, and the other half have the data deleted. Other times, only a few have the data deleted and the others have the desired result.

This is really confusing!!! Can anyone help??

wait a second.....
If i understand it correctly

u have n number of lines like

PS_1 C=C
PS_2 C-C-N
PS_3 C-C=C

and u want to create files like PS_1 which has C=C
and PS_2 which has C-C-N and so on

If thats the case

Code:

cat infile  | perl -e 'while(<>){ chomp ; my ($file_name,$content) = split("\t"); `echo $content > $file_name`;}'

Where infile contains

Code:

cat infile
PS_1    C=C
PS_2    C-C-N
PS_3    C-C=C

If you want to do it using awk

Code:

awk '{print $2 > $1} infile

HTH,
PL

Wow thanks for the fast response. Yes you understood the overall problem correctly, and your solutions seem much simpler even though I don't completely understand all of the commands. I tested the script and it does exactly what you said it would on the test file containing PS_1, PS_2, etc. Unfortunately, the SMILE strings are a little more complicated and include a few characters that seem problematic. This is an example:

PS_1 CCOC(=O)C1C(CC2=C(C(C3=C(C)C=CS3)C(=C(C)N2)C(=O)OCC)C1=O)C4=CC=CC=C4OC

I'll try to modify the script so that these characters (I think its the parentheses) are not problematic. Hopefully I can figure this out. Do you know an easy solution?

Thanks again!

The Problem was

Code:

`echo CCOC(=O)C1C(CC2=C(C(C3=C(C)C=CS3)C(=C(C)N2)C(=O)OCC)C1=O)C4=CC=CC=C4OC > PS_1'

Replace

Code:

echo $content > $file_name

echo "$content" >  $file_name

HTH,
PL

Thanks Daptal! That did exactly what I wanted. Time to figure out what exactly is going with all of the commands.