Shell Programming and Scripting

Hi,

I have an issue with the below script

Code:

nawk 'NR==FNR{a[$1" "$2]=$4" "$5}NR>FNR{print NF?$0:a[x]"\n";if(/^cn:/) x=$0}' FS="[, ]" in1.txt in2.txt > out1.txt

It is taking too long to get a string from in1.text, search for the string in in2.txt and create a new file out1.txt.
Is there any alternative way we can achieve the same faster?

Thanks in advance

Moderator's Comments:

Please use CODE tags rather than ICODE tags when displaying code with more than 1 line or with a single line that might be split due to line length.

NR==FNR doesn't look like the slow part, really. You could put a 'next' in that code block instead of the NR>FNR before the next code block, which may help by skipping that extra check. This doesn't actually look that slow though, honestly.

How large are your input files? How large are your lines? What do they look like? Are they sorted in any fashion?

The code NF?$0:a[x]"\n"; may be creating lots and lots of blank array indexes, a new one for each x not found. If your files are HUGE this may be a significant memory drain. Usually you'd do 'for X in a' before using it. What should it be doing/printing whenever x isn't found in the array?

You could try mawk if you have it, a bytecode-based awk implementation which sometimes has significant performance benefits.