Read data from .csv file through shell script & modify

Shell Programming and Scripting
1

I need to read data from a file called "test.csv" through shell script where the file contains values like name,price,descriptor etc. There are rows where descriptor (& in some rows name) are written as string & other characters like "car_+" OR "bike*" etc where it should contains strings like "car" OR "bike".

Problem is I need to read that file & take out error rows & write those rows into a new file say error.csv & update test.csv with correct rows.

Please let me know how to read & check these things & then update.
OS can be either AIX/HP-UX/Sun-Solaris. Suppose script is bourne shell script.

Thanks,
Raj

2

Post the following:
(1) Your sample input file.
(2) Your desired output.

Give us something to work with.

tyler_durden

3

Alright sample input file : (ofcourse this is not the original file but the original file I need to work is similar to this containing more rows & columns. Original file is office confidential.)

File test.csv :
----------------------------------------------------------------
name price descriptor
Car 50 Mustang
Car_ 70 Audi A5
Bike 30 BMW)
SUV@ 80 Fortuner+_
Car 75 Sierra
--------------------------------
1st row is fine
2nd row -> underscore(car_) need to remove it & put it in a file called error.csv
3rd row -> closing brace ) after BMW & again this row should go to error.csv
4th row -> error at 1st column & last column should go to error.csv
5th row is fine
-----

Now Rows 2, 3 & 4 contain error & they need to be written to error file called error.csv & update test.csv with correct rows. Error rows need not be present in test.csv after these checks.

4

"csv" means "comma separated values". There are no commas in your "test.csv" file, so I'll assume your file looks like the one in the following excerpt -

$
$
$ # show the content of the csv file "f3"
$
$ cat f3
name,price,descriptor
Car,50,Mustang
Car_,70,Audi A5
Bike,30,BMW)
SUV@,80,Fortuner+_
Car,75,Sierra
$
$ # run the awk one-liner to process "f3"
$
$ awk -F, 'NR>1 {if ($1 !~ /^[a-zA-Z0-9 ]+$/ || $3 !~ /^[a-zA-Z0-9 ]+$/){print $0 >"f3.err"} else {print $0 >"f3.tmp"}}' f3 ; mv f3.tmp f3
$
$ # show the content of the error file "f3.err"
$
$ cat f3.err
Car_,70,Audi A5
Bike,30,BMW)
SUV@,80,Fortuner+_
$
$ # show the content of the updated csv file "f3"
$
$ cat f3
Car,50,Mustang
Car,75,Sierra
$
$

tyler_durden