UNIX for Dummies Questions & Answers

Hi,
i am posting this after not finding proper solution in any other discussion threads

I have a text data file with the first row as the column header names.
The file can be comma, tab, space, pipeline, .... delimited

Example of a pipeline delimited:
file1.txt

Code:

colA|colB|colC|colD|colE|colF
1|4|k2|5w|5|7
9|d|78|5|2|9
98|ad|9s|23|0d
j9|3|s9|dw|8k|i

If i call a script say script1.sh and pass the input data delimiter, output delimiter, ColC and ColF as parameters,
it should only display data for those two columns and in the desired format

Note that we should be able to pass any number of column names to the script.
For our example i am using only 2 columns
I don't want to pass column position as parameter because the order of columns can change in the next data file

Code:

cat file1.txt | script1.sh "|" "," ColC ColF

Here "|" means it's the delimiter for the input data, "," is the delimiter to be used for output, the remaining parameters are the column names

The output should be:

Code:

k2,7
78,9
ad,0d
s9,i

It's okay not to print the column header in the output
Any help on how to write script1.sh is greatly appreciated

Thanks,
-srinivas y.

Moderator's Comments:

Please use code tags when posting data and code samples!

Why would you write something that awk already does ?

Do not reinvent the wheel !


Example: to print col4 and col7 if the delimiter of your input is a | and you want the output field separator to be a <tab> :

Code:

awk -F"|" '{print $4 OFS $7}' OFS="\t" yourfile

how about this thread for starters.

another quick example with column 1 and 3 of a test file :

Code:

$ cat tst
1|abc|2
2|def|4
3|ghi|6
4|jkl|8
5|mno|10

Code:

$ awk -F"|" '{print $1 OFS $3}' OFS="\t" tst
1    2
2    4
3    6
4    8
5    10

Code:

$ awk -F"|" '{print $1 OFS $3}' OFS=":" tst
1:2
2:4
3:6
4:8
5:10

ctsgnb, your example has hardcoded column position
I want to able to pass column name as it appears in the column header as parameter
'colD' need not be in the 4position next time.
I think i should have chosen better column names to avoid confusion

---------- Post updated at 02:19 PM ---------- Previous update was at 02:15 PM ----------

vgersh99, thanks but the provided link has an example to search only one column and for a space delimited file
I would like to pass multiple and provide the input and output delimiter

Thanks

You can take the quoted thread as a 'starter/skeleton' and adjust for any additional functionality.

Not sure the scripting language is adapted fo doing such things. Of course it is still possible but more tedious than a select into a Database.

Put your data into a database then

Code:

select colA, colB from yourtable;

there is also a bunch of formatting option available in SQL,SQL+,Pl/Sql so that you can make the rendering fit with your expectations.