Bashscripts.org

Hi All,

Can someone help steer me to the simplest approach to write a script that reads the second column of a csv file and returns the highest value.

example file:

DATE,QUANTITY
Mon Apr 30 01:56:00 2012,42
Mon Apr 30 02:24:00 2012,72
Mon Apr 30 02:56:00 2012,102
Mon Apr 30 02:59:00 2012,132
Mon Apr 30 03:31:00 2012,222
Mon Apr 30 03:38:00 2012,192

TIA

hi,

without using any command, how, step by step, this should be done?

Hmm the above post is a bit useless isn't it.
I'm not experienced yet but in psuedo you have to
navigate to the file
read the first pertinent line
move to the second column after the comma
read that value
send it to where you can work on it -an array I suppose
read the next line's second column
loop through to the end
then read the array and choose the highest value
maybe get the rest of the line data too
send it where you want it

Perhaps in the script a variable might hold the full line and look at the value in a loop and update to the highest after each pass of the loop

A bit mixed up but you will probably get the idea .
I'll bet there are some dam fine solutions on here and I look forward to reading them

no need of an array.
just check if the second column read is greater than the variable that stores greatest value until now.
I didn't put the header into file.in.

Output of this was
# ./readcolumn.sh
./readcolumn.sh: line 5: test: -gt: unary operator expected

worker 10
hirer 15
wages 9
slaves 10
clowns 39
39
[/code]
File was


So I expect its looking at "worker" and saying its not "unary"

add a set -x at the beginning of the script; that will show what's happening.

# ./readcolumn.sh
+ IFS=,
+ read first second rest
+ test -gt 0
./readcolumn.sh: line 7: test: -gt: unary operator expected
+ echo

+ echo store
store
+ IFS=,
+ read first second rest
+ test 10 -gt 0
+ store=10
+ echo worker 10
worker 10
+ echo store 10
store 10
+ IFS=,
+ read first second rest
+ test 15 -gt 10
+ store=' 15'
+ echo hirer 15
hirer 15
+ echo store 15
store 15
+ IFS=,
+ read first second rest
+ test 9 -gt 15
+ echo wages 9
wages 9
+ echo store 15
store 15
+ IFS=,
+ read first second rest
+ test 10 -gt 15
+ echo slaves 10
slaves 10
+ echo store 15
store 15
+ IFS=,
+ read first second rest
+ test 39 -gt 15
+ store=39
+ echo clowns 39
clowns 39
+ echo store 39
store 39
+ IFS=,
+ read first second rest
+ echo store 39
store 39
#

that's it.
if your file might have empty line (which a database shouldn't have), check that line is not empty to process test on values.

Use the correct syntax and it will work as expected.



(( ... )) will only return an error code of 1 if you're trying to process non-integers.

Elegant thanks
could you explain what "read -r _ num _" does?
I assume the "_" moves the columns
How does it know the delimiter to look for ?

EDIT-- IFS is a list of characters so IFS="," does that and _ jumps to the next delimiter

_ is a "dummy variable" that could be any other valid name.
I don't use this character because _ is a special parameter (see man bash /Special Parameters)