UNIX for Advanced & Expert Users

Hi all,

I have a script that receives a file, receives the values and put them on an oracle database.
I need to do this for multiple files that will contain in the end near 500 million records, so performance is a major concern.

until now there was no issue because the file I'm receiving is a csv in the format:

Code:

value1,value2,value3,value4
value1,value2,value3,value4
value1,value2,value3,value4
value1,value2,value3,value4

but management decided to change to a diferent provider and this one is providing the files with the following output:

Code:

field1=value1,field2=value2,field3=value3,field4=value4
field1=value1,field2=value2,field3=value3,field4=value4
field1=value1,field2=value2,field3=value3,field4=value4
field1=value1,field2=value2,field3=value3,field4=value4

previously when i wanted to access a field, for example field 3 i just need to do:

Code:

cut -d"," -f3

I'm trying to get one expression that doesn't require regular expressions to get the same output with the new file but I'm unable to find a way do get this.

Is there any option I can indicate two separators to the cut expression?

Code:

cut -d"," and -d"=" -f6

Using this with both expressions I would get "value3".

In this case I would not require regular expression and just needed to ajust the indexes I use for the several fields...

Thanks in advance.

Best regards,
Ricardo Tomás

Is awk OK?

Code:

awk -F"[,=]" '{print $6}' file

Alternatively you could use two cuts:

Code:

cut -d"," -f3 file | cut -d"=" -f2

I forgot to add something...
I've tried that and this is what i get:

Code:

cut -f1 -d"," test2.txt | cut -f2 -d"="
value1
cut -f2 -d"," test2.txt | cut -f2  -d"="
value2
cut –f3 -d"," test2.txt | cut -f2  -d"="
value3

but I need to get multiple values at once I'm not able to make it work.
Is it possible to configure cut to do that?

For example if I want to get "value1", "value2" and "value3" at once!

It works with AWK. but i would prefer to use cut because cut I know it can be integrated correctly in another tool I use and I'm not sure It as support for AWK.

with AWK:

Code:

awk -F"[,=]" '{print $2 $4 $6 }' test2.txt
value1value2value3

Regards

---------- Post updated at 08:12 AM ---------- Previous update was at 07:49 AM ----------

Got a solution that works for me:

Code:

 
sed s/=/,/g test2.txt | cut -d"," -f2,4,6,8

Thanks.