Looping through a set of files linux based on filepath - linux

Question

I have directory with lots of compressed data with with a couple of file names. I have two file types als.sumstats.lmm.chr and als.sumstats.meta.chr. After chr there is a number 1-22. I want to loop through only the als.sumstats.meta.chr. However, my code is not working. I keep getting gzip: /ALSsummaryGWAS/Summary_Statistics_GWAS_2016/als.sumstats.meta.chr*.txt.gz: no such file or directory, suggesting my files are not being found with my loop. Can someone help. This is what I have right now.
#!/bin/bash
FILES=/ALSsummaryGWAS/Summary_Statistics_GWAS_2016/als.sumstats.meta.chr*.txt.gz
for f in $FILES;
do
echo "$FILES"
echo "extracting columns 2,1,3,9"
gunzip -c $f | awk '{print $2, $1, $3, $14+$15}' >> ALSGWAS.txt
done

Answer 1

In your script snippet, wildcard '*' pattern is stored as a string in the $FILES variable which needs to be evaluated at some point to get the list of matching files.
In order to evaluate it, you can use eval like this:
FILES="ls -1 /ALSsummaryGWAS/Summary_Statistics_GWAS_2016/als.sumstats.meta.chr*.txt.gz"
for f in $(eval $FILES);
do
echo "$FILES"
echo "processing $f"
echo "extracting columns 2,1,3,9"
gunzip -c $f | awk '{print $2, $1, $3, $14+$15}' >> ALSGWAS.txt
done
But eval is not a recommended way to do such operations (eval is dangerous), so you can try it like this:
FILES=$(ls -1 /ALSsummaryGWAS/Summary_Statistics_GWAS_2016/als.sumstats.meta.chr*.txt.gz)
for f in $FILES;
do
echo "$FILES"
echo "processing $f"
echo "extracting columns 2,1,3,9"
gunzip -c $f | awk '{print $2, $1, $3, $14+$15}' >> ALSGWAS.txt
done