UNIX for Dummies Questions & Answers

OK, so I am trying to use awk as a method of accessing a table stored in a file to then provide the capability of a look up table.

The table is stored in a file named "/Users/jhaney/Desktop/assetTypeMapping.tsv" and looks like this:

Code:

aCategory	aLetter	aNumber
AssetCat1	A	123		
AssetCat2	B	124				
AssetCat3	C	125

The file is tab delimited because there is all likelihood that some of the fields will contain space characters and records are separated by UNIX new line characters.

-----------------------
proof of concept:

Code:

#!/bin/sh
#basefilename= echo | basename "$filePath"
luVal="WebAssets"
inputFile="/Users/jhaney/Desktop/assetTypeMapping.tsv"
myResult=`echo | awk -v lookupVal="$luVal" '$1 == lookupVal { print $2 }' "$inputFile"`
echo "$myResult"

This script works great and does exactly what I want it to do.

-----------------------
The problem:

As aforementioned, I would like the freedom of allowing space characters in the data table. However, awk considers spaces to be field separators.

I have seen examples of how to set the field separator:

Code:

awk 'BEGIN { FS = "tab character would go here" } ; { print $2 }'

But I just can't figure out how to combine the two ideas together.

I tried:

Code:

myResult=`echo | awk 'BEGIN { FS = "	" } ; { -v lookupVal="$luVal" $1 == lookupVal { print $2 }}' "$inputFile"`

But I keep getting awk syntax errors no matter how I combine the args.


So, basically, I want to search the first field of each line for "lookupVal" and, when found, return the second (or third, or fourth) field in that record.



Can anyone set me straight?

FS is a variable as any other in awk, so you can set it like any other, e.g. your lookupval. There's several ways to do this (cf man awk):

Code:

awk -v FS="<TAB>" -v lookupVal="$luVal" '...'

or

Code:

awk -v lookupVal="$luVal" 'BEGIN { FS = "<TAB>" } ...'

or

Code:

awk -v lookupVal="$luVal" '... ' FS="<TAB>" ...

. For FS, this one has survived from old days:

Code:

awk -F"<TAB>" ...

What I don't understand is why you echo nothing into a pipe for awk's stdin and then use a file to work upon?

Because I'm a noob. by all means tell me how I ought to be doing it, I want to learn.

Well, first of all, there's man pages and info online documentation that is consulted again and again, even by experienced IT people. Fora like this are willing to give you hints and help when you're stuck.

For above question, consider using either of the examples:

Code:

myResult=$(awk -F"<TAB>" -v lookupVal="$luVal" '$1 == lookupVal { print $2 }' "$inputFile")

<TAB> stands for a single <TAB> character (not sure if ALL awks accept the "\t" notation), and the deprecated backticks have been replaced by the recommended $(...) command substitution syntax.

I did in fact, read the man pages and online examples, but there were so many differences to approach and syntax that I just got pretty mixed up.

Thanks for your assistance.

Note: all awks (except old awk that nobody uses) accept the \t notation, so -F'\t' should work

I just got it to work:

Code:

myResult=$(awk -F"\t" -v lookupVal="$luVal" '$1 == lookupVal { print $2 }' "$inputFile")

I had to use "\t" for the tab character syntax.

Thanks for all your help.

James

Moderator's Comments:

Code tags, please...

---------- Post updated at 03:04 PM ---------- Previous update was at 02:24 PM ----------

I am working OSX. When I create my data table in Excel and save it as a tab delimited file, the default line delimiter is a MAC Carriage return "\r"

I looked through the man page and some other sources on line and I don't see an option to set the record separator "-RS" from the command line the same way we used "-F" to set the field separator.

Here is my working code (Assuming UNIX line feeds)

Code:

assetTypeHandling=$(awk -F"\t" -v lookupVal="$assetType" '$1 == lookupVal { print $2,$3,$4,$5 }' "$assetTypeMapFile")

I tried

Code:

awk -RS"\r" -F"\t" -v ...

and

Code:

awk RS="\r" -F"\t" -v ...

But both of those result in:

Code:

awk: syntax error at source line 1

Is there some way to feed this line of code "\r" as the record separator?