UNIX for Dummies Questions & Answers

does it make any difference in terms of performance while using any of the below mentioned code for the same requirement which processes continuously coming files in the I/P directory .

Please provide ur viewws


ls -tr $SAPRESPONSEGOFILE | sed "s/go/dat/g" |
while read SAPRESPONSEFILES
do

echo "File is ${SAPRESPONSEFILES##*/}"
done

_____________________________________________________
for SAPRESPONSEFILES in `ls $SAPRESPONSEGOFILE | sed "s/go/dat/g"`
do
echo "file is ${SAPRESPONSEFILES##*/}

done
__________________________________________

In your case I don't think it will make any performance difference..
its all depend what you are gonna do inside that loop....

Both technically create the same amount of processes. There may be a small performance increase in the former since it doesn't have to wait for ls to finish completely, but there is a bigger advantage not performance-related: it can handle an arbitrarily large number of files, where the backtick version can only handle the maximum number of arguments allowed by the shell.

Don't know why you have started another thread on this subject. Your original performance issue was about whether sorting by date was a greater performance hit then sorting by name.

Code:

ls -tr $SAPRESPONSEGOFILE | sed "s/go/dat/g" |
while read SAPRESPONSEFILES
do
          echo "File is ${SAPRESPONSEFILES##*/}"
done

As other correspondents remark the "while" construct can deal with any number of files. However, only ONE filename is presented in each iteration:
I still recommend specifying the "-1" parameter (hyphen one).

Code:

ls -1tr $SAPRESPONSEGOFILE | sed "s/go/dat/g" |
while read SAPRESPONSEFILES
do
          echo "File is ${SAPRESPONSEFILES}"
done

If files are continuously arriving in the directory this can create issues with embyonic files (i.e. files which are part-written). Does your process have in intermediate name for files while they are being written?