Count using two txt files - Unix - KSH

[HHG]

Hi,

I have 2 files:-

File A contains a list of numbers but the 1st 7chrs is a unique.

File B contains duplicate sequence number (1st 7chrs)

Any ideas on how I can count how many duplicates in file B.

I suspect I will have to use subscript or substring it but not sure what to use awk/sed or both or maybe somethingelse.

I am using ksh on unix sun solaris v10

if count > 5 (This would be 5 files starting with the same
then
transfer files
else
don't want these files
fi

 

[feherke]

Hi

To find out how many times the first 7 characters are duplicated in file B :

uniq -w 7 -c /file/B

Tested with GNU [tt]uniq[/tt].

You can also do it with [tt]awk[/tt] but it is a bit longer. If you are stronger in [tt]sed[/tt] than Aurélio Marinho Jargas, you can try it in [tt]sed[/tt] too...

Feherke.

http://rootshell.be/~feherke/

 

[HHG]

thanks but I will need to know what the 7 characters are so that I can use it e.g. mv 1234567*.txt destination.
I will have a go will let you know on Monday. thanks again.

 

[columb]

COUNT=0
while read line
do
  short=$(expr "$line" : "\(.......\)*")
  egrep -q ^$short fileA && (( COUNT += 1 ))
done < fileB
if [ $COUNT -gt 5 ]
then 
  do it
else
  dont do it
fi

But there are may other ways

Ceci n'est pas une signature
Columb Healy

 

[columb]

Hmm...
Looks like I misunderstod the question

Ceci n'est pas une signature
Columb Healy

 

[feherke]

Hi

thanks but I will need to know what the 7 characters are so that I can use it e.g. mv 1234567*.txt destination.

And what ? Was not there all the information you needed ?

uniq -w 7 -c /file/B | \
while read nr name; do
  test "$nr" -ge 5 && mv ${name:0:7}*.txt destination/;
done

Feherke.

http://rootshell.be/~feherke/

 

[PHV]

-w is NOT a standard (X/Open or POSIX) option for the uniq command ...

 

[feherke]

Hi

Yep, I had such a feeling. That is why I specified "Tested with GNU [tt]uniq[/tt].".

I bothered to share this solution because HHG still not provided details about his system although was useful in the previous thread too... But I am quite sure it was in vain.

Feherke.

http://rootshell.be/~feherke/

 

[PHV]

HHG still not provided details about his system
Really ?

am using ksh on unix sun solaris v10

 

[feherke]

Hi

Oops. I missed that line. No idea how. Sorry.

Feherke.

http://rootshell.be/~feherke/

 

[HHG]

Hi thanks for the tips guys.
I'm back on the case now. I have done the following

cat xfer_list2.txt | awk '{print substr($1,1,7)}' | uniq -c

this gives me the following output two fields :-

Fld1 Field2
6 0000016
6 0000017
6 0000018
5 0000019
6 0000020
5 0000021
6 0000022
6 0000024
1 EOF2813
2 xfer_li

Any ideas how I might do test feild1 to be greater than or equal to 5 and do an action on the second field. e.g. on the first line of my example above field1 = 6, field2=0000016

I want to be able to do mv field2* to destination
but where its less then 5 I want to move those to another destination.

any help you can give me would be most appreciated, I really need your help to finish this piece of work as I am not very strong on unix scripting. By the way I am a her not a him, just so you know. Thanks

 

[feherke]

Hi

You can combine that with my earlier code :

[red]awk '{print substr($1,1,7)}' xfer_list2.txt | uniq -c[/red] | \
while read nr name; do
  test "$nr" -ge 5 && mv $name*.txt destination/
done

By the way, avoid UUOC.

Feherke.

http://rootshell.be/~feherke/

 

[HHG]

Actually, you feherke you already give me the answer in the while read nr name statement, which I have done and works fine.
Thanks for your help, you saved the day.

 

[HHG]

Just one other question, as there is not an else statement. How do you think I can move the ones that are left over ie the ones less then 4. Hmmm, maybe I do the same statement again but do a < 5 though I am not sure of the syntax for less than.

 

[HHG]

Ok I found it in my book. I used the same statement with but instead of -ge I have used -le

Thanks all again for your help.

 

[PHV]

I'd use -lt instead of -le (unless you want some error for the files with a count=5 ...)

 

[feherke]

Hi

If you need an else branch then better change to a proper [tt]if[/tt] statement. The [tt]&&[/tt] are nice by alone but tend to become cryptic if there are many. And for the two [tt]test[/tt] commands two expressions will need to be evaluated, while the [tt]if[/tt] would have only one.

awk '{print substr($1,1,7)}' xfer_list2.txt | uniq -c | \
while read nr name; do
  [red]if [[/red] "$nr" -ge 5 [red]]; then[/red]
    mv $name*.txt destination/
  [red]else[/red]
    [gray]# echo "too few $name* files. all $n skipped." >> log_file[/gray]
  [red]fi[/red]
done

Feherke.

http://rootshell.be/~feherke/

 

[HHG]

Hi PHV

I have done the following to take into account = 5

cat xfer_list2.txt | awk '{print substr($1,1,7)}' | uniq -c |\
while read nr name
do
test "$nr" -le 4 && mv ${name}* /l/test/var/pmcli/ifc/cru/export/prexfr
done

Just about to test the script.

 

[HHG]

Ahhh I like that one Feherke. I might just give that a go, but might have to tweak it alittle so I can move those files into another area but can use the error trapping you have given me. Thanks your brill!

 

[PHV]

Have a look here:

http://uuoc.com/def.php