Awk, matching records in a comma delimited file

Ronnie Your
#1 / 22

matching records in a comma delimited file

Hi,

I have a comma delimited file as

REG,12619,GITT,LOM,52891
GTU,15679211,GTSP,LOSN,82248
TAP1132,16791,JSAEDJH,SSOPA.,129225
SR,17551,AAFSSSA,89114120,AAU
FEP,22576,SSJ,222,DEP

The above file would have around 8-10 million records.

There is another file which has the records based on which I need to filter
the above file.
This will have around 600-700 records

REG
E
SR
SPSDDS
WSH
DSSL

I tried grep -f as
grep -f smallfile.txt bigfile.txt > filteredrecords.txt

But the problem is that grep matches the whole record for occurrences of the
string.

I would like the records in the small file to match the values in the first
field of the big file.

Hence I would want the result to be
REG,12619,GITT,LOM,52891
SR,17551,AAFSSSA,89114120,AAU

But grep gives me

REG,12619,GITT,LOM,52891
TAP1132,16791,JSAEDJH,SSOPA.,129225
FEP,22576,SSJ,222,DEP

Any pointers how do I go about doing this.

Thanks
Ronnie

Fri, 07 Jan 2005 21:07:06 GMT

Torsten Kirschne
#2 / 22

matching records in a comma delimited file

Quote:

> Hi,

> I have a comma delimited file as

> REG,12619,GITT,LOM,52891
> GTU,15679211,GTSP,LOSN,82248
> TAP1132,16791,JSAEDJH,SSOPA.,129225
> SR,17551,AAFSSSA,89114120,AAU
> FEP,22576,SSJ,222,DEP

> The above file would have around 8-10 million records.

> There is another file which has the records based on which I need to
> filter the above file.
> This will have around 600-700 records

> REG
> E
> SR
> SPSDDS
> WSH
> DSSL

> I tried grep -f as
> grep -f smallfile.txt bigfile.txt > filteredrecords.txt

> But the problem is that grep matches the whole record for occurrences of
> the string.

> I would like the records in the small file to match the values in the
> first field of the big file.

> Hence I would want the result to be
> REG,12619,GITT,LOM,52891
> SR,17551,AAFSSSA,89114120,AAU

> But grep gives me

> REG,12619,GITT,LOM,52891
> TAP1132,16791,JSAEDJH,SSOPA.,129225
> FEP,22576,SSJ,222,DEP

Maybe this helps...

1. sort datafile > data.sorted
2. sort patternfile > pattern.sorted
3. join -1 1 -2 1 -t, data.sorted pattern.sorted

I am somewhat uncertain about whether join works with
giant files like Yours.

Other than join, nawk, and perl come to mind.

Fri, 07 Jan 2005 21:54:17 GMT

Robert Stearn
#3 / 22

matching records in a comma delimited file

A simple awk script (possibly space consumptive, depending on your awk
version):
awk 'NR==FNR { wanted[$1]=1; next }; $1 in wanted { print $0)' \
keyfile datafile > outfile

This is untested code, proceed with caution.

Quote:

> Hi,

> I have a comma delimited file as

> REG,12619,GITT,LOM,52891
> GTU,15679211,GTSP,LOSN,82248
> TAP1132,16791,JSAEDJH,SSOPA.,129225
> SR,17551,AAFSSSA,89114120,AAU
> FEP,22576,SSJ,222,DEP

> The above file would have around 8-10 million records.

> There is another file which has the records based on which I need to filter
> the above file.
> This will have around 600-700 records

> REG
> E
> SR
> SPSDDS
> WSH
> DSSL

> I tried grep -f as
> grep -f smallfile.txt bigfile.txt > filteredrecords.txt

> But the problem is that grep matches the whole record for occurrences of the
> string.

> I would like the records in the small file to match the values in the first
> field of the big file.

> Hence I would want the result to be
> REG,12619,GITT,LOM,52891
> SR,17551,AAFSSSA,89114120,AAU

> But grep gives me

> REG,12619,GITT,LOM,52891
> TAP1132,16791,JSAEDJH,SSOPA.,129225
> FEP,22576,SSJ,222,DEP

> Any pointers how do I go about doing this.

> Thanks
> Ronnie

Fri, 07 Jan 2005 22:09:17 GMT

Dan Haygoo
#4 / 22

matching records in a comma delimited file

Quote:

> Hi,

> I have a comma delimited file as

> REG,12619,GITT,LOM,52891
> GTU,15679211,GTSP,LOSN,82248
> TAP1132,16791,JSAEDJH,SSOPA.,129225
> SR,17551,AAFSSSA,89114120,AAU
> FEP,22576,SSJ,222,DEP

> The above file would have around 8-10 million records.

> There is another file which has the records based on which I need to
filter
> the above file.
> This will have around 600-700 records

> REG
> E
> SR
> SPSDDS
> WSH
> DSSL

> I tried grep -f as
> grep -f smallfile.txt bigfile.txt > filteredrecords.txt

> But the problem is that grep matches the whole record for occurrences of
the
> string.

> I would like the records in the small file to match the values in the
first
> field of the big file.

> Hence I would want the result to be
> REG,12619,GITT,LOM,52891
> SR,17551,AAFSSSA,89114120,AAU

> But grep gives me

> REG,12619,GITT,LOM,52891
> TAP1132,16791,JSAEDJH,SSOPA.,129225
> FEP,22576,SSJ,222,DEP

> Any pointers how do I go about doing this.

> Thanks
> Ronnie

Ronnie -

Look at join if you you don't want any fancy processing. If you want to do
some nifty statistics couting or field calculation, then awk might be
better.

Here are a couple of awk solutions: First, you can read the keys into an
array, and then, for each data line, see if the key is in the array. Use FS
to acheive the comma-delimiting. This solution is kind of bad, since, for
every line, you have to search the whole array of keys to find a match. If
you have n data lines, and awk just happens to store the array indicies in a
tree, this ends up being an n*log(m) solution.

Second, you can sort the data, and then walk through the data and keys
together, require neither pre-reading and storage, nor a search through a
list, making this an n+m solution (after sorting, of course). This way does
give sorted output.

- Dan

join1.awk
----------
BEGIN {
FS = ","
keyFile = ARGV[1]
delete ARGV[1]
while ( (ok = (getline < keyFile)) > 0 )
keys[$0]
close(keyFile)
if (ok < 0) { print "keyFile error"; exit }
}
$1 in keys { print }

keys
----------
DSSL
E
REG
SPSDDS
SR
WSH

data
----------
FEP,22576,SSJ,222,DEP
GTU,15679211,GTSP,LOSN,82248
REG,12619,GITT,LOM,52891
SR,17551,AAFSSSA,89114120,AAU
TAP1132,16791,JSAEDJH,SSOPA.,129225

(output)
----------
>awk -f join1.awk keys data
REG,12619,GITT,LOM,52891
SR,17551,AAFSSSA,89114120,AAU

join2.awk
----------
BEGIN {
FS = ","
keyFile = ARGV[1]
delete ARGV[1]
}

$1 > key {
while ( (ok = (getline key < keyFile)) > 0 && key < $1 ) ;
if (ok < 0) { print "keyFile error"; exit }
}

$1 == key { print }

END { if (!ok) print "getline error" }

(output) (gawk/Win98)
----------
>sort keys > keys.srt
>type keys.srt
DSSL
E
REG
SPSDDS
SR
WSH

>sort data > data.srt
>type data.srt
FEP,22576,SSJ,222,DEP
GTU,15679211,GTSP,LOSN,82248
REG,12619,GITT,LOM,52891
SR,17551,AAFSSSA,89114120,AAU
TAP1132,16791,JSAEDJH,SSOPA.,129225

>awk -f join2.awk keys.srt data.srt
REG,12619,GITT,LOM,52891
SR,17551,AAFSSSA,89114120,AAU

Fri, 07 Jan 2005 22:15:14 GMT

Kenny McCorma
#5 / 22

matching records in a comma delimited file

Quote:

>Hi,

>I have a comma delimited file as

>REG,12619,GITT,LOM,52891
>GTU,15679211,GTSP,LOSN,82248
>TAP1132,16791,JSAEDJH,SSOPA.,129225
>SR,17551,AAFSSSA,89114120,AAU
>FEP,22576,SSJ,222,DEP

>The above file would have around 8-10 million records.

OK - let's assume away the usual "parsing CSV" issues (that is, we assume
that your file has no embedded commas or quotes or other funny stuff.

Quote:

>There is another file which has the records based on which I need to filter
>the above file.
>This will have around 600-700 records

>REG
>E
>SR
>SPSDDS
>WSH
>DSSL

>I tried grep -f as
>grep -f smallfile.txt bigfile.txt > filteredrecords.txt

gawk -F, 'ARGIND == 1 { x[$0];next } $1 in x' smallfile.txt bigfile.txt > filteredrecords.txt

Fri, 07 Jan 2005 22:32:24 GMT

Dan Haygoo
#6 / 22

matching records in a comma delimited file

Quote:

> A simple awk script (possibly space consumptive, depending on your awk
> version):
> awk 'NR==FNR { wanted[$1]=1; next }; $1 in wanted { print $0)' \
> keyfile datafile > outfile

> This is untested code, proceed with caution.

> > Hi,

> > I have a comma delimited file as

> > REG,12619,GITT,LOM,52891
> > GTU,15679211,GTSP,LOSN,82248
> > TAP1132,16791,JSAEDJH,SSOPA.,129225
> > SR,17551,AAFSSSA,89114120,AAU
> > FEP,22576,SSJ,222,DEP

> > The above file would have around 8-10 million records.

> > There is another file which has the records based on which I need to
filter
> > the above file.
> > This will have around 600-700 records

> > REG
> > E
> > SR
> > SPSDDS
> > WSH
> > DSSL

> > I tried grep -f as
> > grep -f smallfile.txt bigfile.txt > filteredrecords.txt

> > But the problem is that grep matches the whole record for occurrences of
the
> > string.

> > I would like the records in the small file to match the values in the
first
> > field of the big file.

> > Hence I would want the result to be
> > REG,12619,GITT,LOM,52891
> > SR,17551,AAFSSSA,89114120,AAU

> > But grep gives me

> > REG,12619,GITT,LOM,52891
> > TAP1132,16791,JSAEDJH,SSOPA.,129225
> > FEP,22576,SSJ,222,DEP

> > Any pointers how do I go about doing this.

> > Thanks
> > Ronnie

Ronnie, Robert -

Somewhere in here, you'll want the parameter "-F," to awk, or use 'BEGIN {FS
= ","}' in your code. Some other general notes:
- "NR == NFR/next" clauses are often more effcient if done in BEGIN with
getline. But a good idea to shorten this for the command line!
- For the "in" test to work, awk doesn't require the setting of the value;
just the reference,
so in this case, "wanted[$1]" is equivalent to "wanted[$1]=1"
- $0 is the default arg to "print", so "{print}" is equivalent to "{print
$0)" [sic].
- And, even though I used the "{print}" construct in my other post, the
default action of a pattern is "{print}", so that can be eliminated.

To save typing, I shortened "wanted" to "k". This tested output
(gawk/Win98) uses data I posted elsewhere in this thread.

- Dan

(output)
----------
>awk -F, "NR==FNR{k[$1]; next} $1 in k" keys data
REG,12619,GITT,LOM,52891
SR,17551,AAFSSSA,89114120,AAU

Fri, 07 Jan 2005 22:40:40 GMT

William Par
#7 / 22

matching records in a comma delimited file

Quote:

> Hi,

> I have a comma delimited file as

> REG,12619,GITT,LOM,52891
> GTU,15679211,GTSP,LOSN,82248
> TAP1132,16791,JSAEDJH,SSOPA.,129225
> SR,17551,AAFSSSA,89114120,AAU
> FEP,22576,SSJ,222,DEP

> The above file would have around 8-10 million records.

> There is another file which has the records based on which I need to filter
> the above file.
> This will have around 600-700 records

> REG
> E
> SR
> SPSDDS
> WSH
> DSSL

> I tried grep -f as
> grep -f smallfile.txt bigfile.txt > filteredrecords.txt

> But the problem is that grep matches the whole record for occurrences of the
> string.

> I would like the records in the small file to match the values in the first
> field of the big file.

> Hence I would want the result to be
> REG,12619,GITT,LOM,52891
> SR,17551,AAFSSSA,89114120,AAU

> But grep gives me

> REG,12619,GITT,LOM,52891
> TAP1132,16791,JSAEDJH,SSOPA.,129225
> FEP,22576,SSJ,222,DEP

> Any pointers how do I go about doing this.

> Thanks
> Ronnie

Convert the search pattern to
^...,
format,
sed 's/.*/^&,/' smallfile.txt > s
grep -f s bigfile.txt > filteredrecords.txt

--

8-CPU Cluster, Hosting, NAS, Linux, LaTeX, python, vim, mutt, tin

Fri, 07 Jan 2005 22:56:37 GMT

Robert Stearn
#8 / 22

matching records in a comma delimited file

I made an obvious error. I left out ' -F, ' in the awk statement. It
should immediately follow the 'awk', without the apostrophes.

Quote:

> A simple awk script (possibly space consumptive, depending on your awk
> version):
> awk 'NR==FNR { wanted[$1]=1; next }; $1 in wanted { print $0)' \
> keyfile datafile > outfile

> This is untested code, proceed with caution.

> > Hi,

> > I have a comma delimited file as

> > REG,12619,GITT,LOM,52891
> > GTU,15679211,GTSP,LOSN,82248
> > TAP1132,16791,JSAEDJH,SSOPA.,129225
> > SR,17551,AAFSSSA,89114120,AAU
> > FEP,22576,SSJ,222,DEP

> > The above file would have around 8-10 million records.

> > There is another file which has the records based on which I need to filter
> > the above file.
> > This will have around 600-700 records

> > REG
> > E
> > SR
> > SPSDDS
> > WSH
> > DSSL

> > I tried grep -f as
> > grep -f smallfile.txt bigfile.txt > filteredrecords.txt

> > But the problem is that grep matches the whole record for occurrences of the
> > string.

> > I would like the records in the small file to match the values in the first
> > field of the big file.

> > Hence I would want the result to be
> > REG,12619,GITT,LOM,52891
> > SR,17551,AAFSSSA,89114120,AAU

> > But grep gives me

> > REG,12619,GITT,LOM,52891
> > TAP1132,16791,JSAEDJH,SSOPA.,129225
> > FEP,22576,SSJ,222,DEP

> > Any pointers how do I go about doing this.

> > Thanks
> > Ronnie

Fri, 07 Jan 2005 23:13:24 GMT

Ronnie Your
#9 / 22

matching records in a comma delimited file

I tried the Join method and it seems to work fine for most part of it.

But join doesnt seem to work for where the patterns are
CHAR, CNPZY

Lets take this example where the records in the data file are

CWEI,1995,NORAM,SECNRPS, -30.350000, -30.350000,YR
CNPZY,1995,CON,SECPDFNC, 1340.277644, 1340.947783,YR
CHAR,1995,US,UDGUUT, 2500.000000, 2500.000000,YR
HHL,1995,WOUSCON,RRCAPREA, -17714.285714, -17714.285714,YR
COHO,1995,REGROLL,UDNPOGDW, 6.400000, 6.400000,YR

The pattern file will have records as

CWEI
CNPZY
CHAR
HHL
COHO

Now the join
join -1 1 -2 1 -t, data pattern

gives the output as

CWEI,1995,NORAM,SECNRPS, -30.350000, -30.350000,YR
HHL,1995,WOUSCON,RRCAPREA, -17714.285714, -17714.285714,YR
COHO,1995,REGROLL,UDNPOGDW, 6.400000, 6.400000,YR

It has omitted the lines with CWEI and CNPZY as the first field.

Is there anything I am missing here.

Also when I sort
sort bigfile > bigsortedfile
It gives me a warning
Warning: use of redirected output for large sorts is usually slower than
specifying the output file name directly to
sort.

Please suggest.

Thanks
Ronnie

Quote:

> > Hi,

> > I have a comma delimited file as

> > REG,12619,GITT,LOM,52891
> > GTU,15679211,GTSP,LOSN,82248
> > TAP1132,16791,JSAEDJH,SSOPA.,129225
> > SR,17551,AAFSSSA,89114120,AAU
> > FEP,22576,SSJ,222,DEP

> > The above file would have around 8-10 million records.

> > There is another file which has the records based on which I need to
> > filter the above file.
> > This will have around 600-700 records

> > REG
> > E
> > SR
> > SPSDDS
> > WSH
> > DSSL

> > I tried grep -f as
> > grep -f smallfile.txt bigfile.txt > filteredrecords.txt

> > But the problem is that grep matches the whole record for occurrences of
> > the string.

> > I would like the records in the small file to match the values in the
> > first field of the big file.

> > Hence I would want the result to be
> > REG,12619,GITT,LOM,52891
> > SR,17551,AAFSSSA,89114120,AAU

> > But grep gives me

> > REG,12619,GITT,LOM,52891
> > TAP1132,16791,JSAEDJH,SSOPA.,129225
> > FEP,22576,SSJ,222,DEP

> Maybe this helps...

> 1. sort datafile > data.sorted
> 2. sort patternfile > pattern.sorted
> 3. join -1 1 -2 1 -t, data.sorted pattern.sorted

> I am somewhat uncertain about whether join works with
> giant files like Yours.

> Other than join, nawk, and perl come to mind.

Sat, 08 Jan 2005 00:27:38 GMT

Kenny McCorma
#10 / 22

matching records in a comma delimited file

Quote:

>I tried the Join method and it seems to work fine for most part of it.

>But join doesnt seem to work for where the patterns are
>CHAR, CNPZY

Why are you doing this the hard way, with sort, join, etc, etc when I and
others have posted AWK solutions?

Which NG are you following this in? (I.e., if you weren't interested in
AWK solutions, why x-post to comp.lang.awk?)

Sat, 08 Jan 2005 01:11:09 GMT

Ronnie Your
#11 / 22

matching records in a comma delimited file

Hi Kenny,
This solution works great for me.
gawk -F, 'ARGIND == 1 { x[$0];next } $1 in x' smallfile.txt bigfile.txt >
filteredrecords.txt

Can you please explain what steps it goes through so that i understand it
better.

Thanks a Lot
Ronnie

Quote:

> >Hi,

> >I have a comma delimited file as

> >REG,12619,GITT,LOM,52891
> >GTU,15679211,GTSP,LOSN,82248
> >TAP1132,16791,JSAEDJH,SSOPA.,129225
> >SR,17551,AAFSSSA,89114120,AAU
> >FEP,22576,SSJ,222,DEP

> >The above file would have around 8-10 million records.

> OK - let's assume away the usual "parsing CSV" issues (that is, we assume
> that your file has no embedded commas or quotes or other funny stuff.

> >There is another file which has the records based on which I need to
filter
> >the above file.
> >This will have around 600-700 records

> >REG
> >E
> >SR
> >SPSDDS
> >WSH
> >DSSL

> >I tried grep -f as
> >grep -f smallfile.txt bigfile.txt > filteredrecords.txt

> gawk -F, 'ARGIND == 1 { x[$0];next } $1 in x' smallfile.txt bigfile.txt >
filteredrecords.txt

Sat, 08 Jan 2005 05:03:10 GMT

John W. Krah
#12 / 22

matching records in a comma delimited file

Quote:

> Hi,

> I have a comma delimited file as

> REG,12619,GITT,LOM,52891
> GTU,15679211,GTSP,LOSN,82248
> TAP1132,16791,JSAEDJH,SSOPA.,129225
> SR,17551,AAFSSSA,89114120,AAU
> FEP,22576,SSJ,222,DEP

> The above file would have around 8-10 million records.

> There is another file which has the records based on which I need to filter
> the above file.
> This will have around 600-700 records

> REG
> E
> SR
> SPSDDS
> WSH
> DSSL

> I tried grep -f as
> grep -f smallfile.txt bigfile.txt > filteredrecords.txt

> But the problem is that grep matches the whole record for occurrences of the
> string.

> I would like the records in the small file to match the values in the first
> field of the big file.

> Hence I would want the result to be
> REG,12619,GITT,LOM,52891
> SR,17551,AAFSSSA,89114120,AAU

> But grep gives me

> REG,12619,GITT,LOM,52891
> TAP1132,16791,JSAEDJH,SSOPA.,129225
> FEP,22576,SSJ,222,DEP

John
--
use Perl;
program
fulfillment

Sat, 08 Jan 2005 05:27:04 GMT

Dan Haygoo
#13 / 22

matching records in a comma delimited file

Quote:

> But join doesnt seem to work for where the patterns are
> CHAR, CNPZY

I imagine your data is messed up then--look for embedded control characters
with a hex editor. Can you think of any other reasonable explanation why,
of 10 million records, joined to 700 keys, only a few odd cases would
misbehave? Use one of the many awk solutions, or even the perl solution
(that looks suspect to me, because the poster didn't indicate any testing),
and see if you get the same results.

- Dan

Quote:

> I tried the Join method and it seems to work fine for most part of it.

> But join doesnt seem to work for where the patterns are
> CHAR, CNPZY

> Lets take this example where the records in the data file are

> CWEI,1995,NORAM,SECNRPS, -30.350000, -30.350000,YR
> CNPZY,1995,CON,SECPDFNC, 1340.277644, 1340.947783,YR
> CHAR,1995,US,UDGUUT, 2500.000000, 2500.000000,YR
> HHL,1995,WOUSCON,RRCAPREA, -17714.285714, -17714.285714,YR
> COHO,1995,REGROLL,UDNPOGDW, 6.400000, 6.400000,YR

> The pattern file will have records as

> CWEI
> CNPZY
> CHAR
> HHL
> COHO

> Now the join
> join -1 1 -2 1 -t, data pattern

> gives the output as

> CWEI,1995,NORAM,SECNRPS, -30.350000, -30.350000,YR
> HHL,1995,WOUSCON,RRCAPREA, -17714.285714, -17714.285714,YR
> COHO,1995,REGROLL,UDNPOGDW, 6.400000, 6.400000,YR

> It has omitted the lines with CWEI and CNPZY as the first field.

> Is there anything I am missing here.

> Also when I sort
> sort bigfile > bigsortedfile
> It gives me a warning
> Warning: use of redirected output for large sorts is usually slower than
> specifying the output file name directly to
> sort.

> Please suggest.

> Thanks
> Ronnie

> > > Hi,

> > > I have a comma delimited file as

> > > REG,12619,GITT,LOM,52891
> > > GTU,15679211,GTSP,LOSN,82248
> > > TAP1132,16791,JSAEDJH,SSOPA.,129225
> > > SR,17551,AAFSSSA,89114120,AAU
> > > FEP,22576,SSJ,222,DEP

> > > The above file would have around 8-10 million records.

> > > There is another file which has the records based on which I need to
> > > filter the above file.
> > > This will have around 600-700 records

> > > REG
> > > E
> > > SR
> > > SPSDDS
> > > WSH
> > > DSSL

> > > I tried grep -f as
> > > grep -f smallfile.txt bigfile.txt > filteredrecords.txt

> > > But the problem is that grep matches the whole record for occurrences
of
> > > the string.

> > > I would like the records in the small file to match the values in the
> > > first field of the big file.

> > > Hence I would want the result to be
> > > REG,12619,GITT,LOM,52891
> > > SR,17551,AAFSSSA,89114120,AAU

> > > But grep gives me

> > > REG,12619,GITT,LOM,52891
> > > TAP1132,16791,JSAEDJH,SSOPA.,129225
> > > FEP,22576,SSJ,222,DEP

> > Maybe this helps...

> > 1. sort datafile > data.sorted
> > 2. sort patternfile > pattern.sorted
> > 3. join -1 1 -2 1 -t, data.sorted pattern.sorted

> > I am somewhat uncertain about whether join works with
> > giant files like Yours.

> > Other than join, nawk, and perl come to mind.

Sat, 08 Jan 2005 11:11:10 GMT

Ronnie Your
#14 / 22

matching records in a comma delimited file

Hi Guys,
This solution works for me

gawk -F, 'ARGIND == 1 { x[$0];next } $1 in x' smallfile.txt bigfile.txt >
filteredrecords.txt

But I am not sure what exactly it does .I am a newbee and even though I
tried looking up in tutorials / manuals to understand what exactly is
happenning in the above statement I couldnt figure it out.

Can you pleaes explain it.

Thanks a lot
Ronnie

Sat, 08 Jan 2005 23:55:51 GMT

Robert Stearn
#15 / 22

matching records in a comma delimited file

I'll try. It appears to use one gawk specific feature, which I am
unfamiliar with and uses a lot of short hand:

Quote:

> Hi Guys,
> This solution works for me
> gawk -F, 'ARGIND == 1 { x[$0];next } $1 in x' smallfile.txt bigfile.txt

111 22222222222 33333 4444
5555555
5 for records in the second file,
if
their first field is among the
subscripts of x, take the
default
action (elided) of printing the
record
4 go to the next record in the input
3 use the record content as a subscript for the
array x;
the unmentioned value of x[$0] is ""
2 if we are processing the first file mentioned in the
argument list;
this is the gawk extension I mentioned; for standard awk
use NR==FNR
1 set the field separator to ','

Hope that helps.

Quote:

> filteredrecords.txt

> But I am not sure what exactly it does .I am a newbee and even though I
> tried looking up in tutorials / manuals to understand what exactly is
> happenning in the above statement I couldnt figure it out.

> Can you pleaes explain it.

> Thanks a lot
> Ronnie

Sun, 09 Jan 2005 00:24:09 GMT