regex question

I need to find patterns like these (e.g. with sed or perl or grep):

G1150G111
00443E104

etc. That is, 9 digit words made only of letters or digits, of which at
least one character is a digit. The letters can occur in random positions.

What would be the pattern to match? Thanks!

In article <pan.2008.05.25.17.19.45@verizon.net>,
Amadeus W.M. wrote:

> I need to find patterns like these (e.g. with sed or perl or grep):
>
> G1150G111
> 00443E104
>
> etc. That is, 9 digit words made only of letters or digits, of which
> at least one character is a digit. The letters can occur in random
> positions.

perl -ne '
while (/(^|[^[:alnum:]])([[:alnum:]]{9})([^[:alnum:]]|$)/g) {
if ($2 =~ /[[:digit:]]/) {print;last;}
}' <infile >outfile


Regards,
Marcel

--
printf -v email $(echo \ 155 141 162 143 145 154 155 141 162 \
143 145 154 100 157 162 141 156 147 145 56 156 154 | tr \ \\)
# O Herr, lass Hirn vom Himmel fallen! #

On Mon, 26 May 2008 00:10:08 +0200, Marcel Bruinsma wrote:

> In article <pan.2008.05.25.17.19.45@verizon.net>,
> Amadeus W.M. wrote:
>
>> I need to find patterns like these (e.g. with sed or perl or grep):
>>
>> G1150G111
>> 00443E104
>>
>> etc. That is, 9 digit words made only of letters or digits, of which at
>> least one character is a digit. The letters can occur in random
>> positions.
>
> perl -ne '
> while (/(^|[^[:alnum:]])([[:alnum:]]{9})([^[:alnum:]]|$)/g) {
> if ($2 =~ /[[:digit:]]/) {print;last;}
> }' <infile >outfile
>
>
> Regards,
> Marcel

Thanks! I'm not sure this will work for what I need though. Perhaps my
initial question was incomplete. I have a file with many lines of the form


company type companyId $amount #shares etc.

For instance:

GENERAL MTRS CORP Preferred 370442691 4,602 200,000
Shrs Shared-Defined 1 200,000


The file has many lines like this, but not only. I'm trying to find the
lines of this form, and within each line found, extract the companyId,
$amount and #shares.

To thie ens, I'm searching for the pattern "companyId $amount #shares". I
have something like

([[:alnum:]]{9})\s+(\$?number_pattern)\s+(number_pattern)

where number_pattern is something that matches numbers, with or without
commas. With [[:alnum:]]{9} for companyId, followed by the other two
patterns, in the example above I pick up

companyId = Preferred
$amount = 370442691
#shares = 4,602

which would be wrong (but the program thinks it's ok). I need to change
the companyId pattern from the simple minded [[:alnum:]]{9} to something
to include at least 1 digit. And keep the next two patterns.

On Mon, 26 May 2008 00:10:08 +0200, Marcel Bruinsma wrote:

> In article <pan.2008.05.25.17.19.45@verizon.net>,
> Amadeus W.M. wrote:
>
>> I need to find patterns like these (e.g. with sed or perl or grep):
>>
>> G1150G111
>> 00443E104
>>
>> etc. That is, 9 digit words made only of letters or digits, of which at
>> least one character is a digit. The letters can occur in random
>> positions.
>
> perl -ne '
> while (/(^|[^[:alnum:]])([[:alnum:]]{9})([^[:alnum:]]|$)/g) {
> if ($2 =~ /[[:digit:]]/) {print;last;}
> }' <infile >outfile
>
>
> Regards,
> Marcel

I guess I need something like

([[:digit:]][[:alnum:]]{8})|([[:alnum:]]{1}[[:digit:]][[:alnum:]]{7})| etc.

That is, keep moving the [[:digit:]] over each of the 9 possible
positions. Is there a smarter way to write this?

In article <pan.2008.05.26.02.15.07@verizon.net>,
Amadeus W.M. wrote:

>>
>>> I need to find patterns like these (e.g. with sed or perl or grep):
>>>
>>> G1150G111
>>> 00443E104
>>>
>> perl -ne '
>> while (/(^|[^[:alnum:]])([[:alnum:]]{9})([^[:alnum:]]|$)/g) {
>> if ($2 =~ /[[:digit:]]/) {print;last;}
>> }' <infile >outfile
>
> Thanks! I'm not sure this will work for what I need though. Perhaps my
> initial question was incomplete. I have a file with many lines of the
> form
>
> company type companyId $amount #shares etc.
>
> For instance:
>
> GENERAL MTRS CORP Preferred 370442691 4,602 200,000
> Shrs Shared-Defined 1 200,000
>
>
> The file has many lines like this, but not only. I'm trying to find
> the lines of this form, and within each line found, extract the
> companyId, $amount and #shares.
>
> To thie ens, I'm searching for the pattern "companyId $amount
> #shares". I have something like
>
> ([[:alnum:]]{9})\s+(\$?number_pattern)\s+(number_pattern)
>
> where number_pattern is something that matches numbers, with or
> without commas. With [[:alnum:]]{9} for companyId, followed by the
> other two patterns, in the example above I pick up
>
> companyId = Preferred
> $amount = 370442691
> #shares = 4,602
>
> which would be wrong (but the program thinks it's ok). I need to
> change the companyId pattern from the simple minded [[:alnum:]]{9} to
> something to include at least 1 digit. And keep the next two patterns.

The '$2 =~ /[[:digit:]]/' is the check for 'at least one digit', but
only one expression is also possible, just more complicated:

#!/usr/bin/perl
$p =
'[[:blank:]]([[:digit:]][[:alnum:]]{8}|'
..'[[:alnum:]][[:digit:]][[:alnum:]]{7}|'
..'[[:alnum:]]{2}[[:digit:]][[:alnum:]]{6}|'
..'[[:alnum:]]{3}[[:digit:]][[:alnum:]]{5}|'
..'[[:alnum:]]{4}[[:digit:]][[:alnum:]]{4}|'
..'[[:alnum:]]{5}[[:digit:]][[:alnum:]]{3}|'
..'[[:alnum:]]{6}[[:digit:]][[:alnum:]]{2}|'
..'[[:alnum:]]{7}[[:digit:]][[:alnum:]]|'
..'[[:alnum:]]{8}[[:digit:]])'
..'[[:blank:]]+([[:digit:]]+,[[:digit:]]+)'
..'[[:blank:]]+([[:digit:]]+,[[:digit:]]+)';
while (<DATA>) {
if (/$p/o) {
$companyID = $1;
$amount = $2;
$shares = $3;
print "|$companyID|$amount|$shares|\n";
}
}
__END__
GENERAL MTRS CORP Preferred 370442691 4,602 200,000
Shrs Shared-Defined 1 200,000

If the companyIDs contain no lowercase, you should
replace all '[:alnum:]' by '[:upper:][:digit:]'.


Regards,
Marcel

--
printf -v email $(echo \ 155 141 162 143 145 154 155 141 162 \
143 145 154 100 157 162 141 156 147 145 56 156 154 | tr \ \\)
# O Herr, lass Hirn vom Himmel fallen! #