[SGVLUG] Grep and the command line...

[Emerson, Tom (*IC)]

Actually, I want to find where a quote immediately follows a tab -- the
"normal" case is that it doesn't, so "tab only" searching would find
every line in the file.

In any case, the real difficulty comes in specifying the quote character
-- remember, this is being done in DOS, not bash/ksh/csh/whatever-sh.
When I type in this:

C:\> grep /"/ CV5055

And then press <tab> [to auto-complete the filename] it changes to this:

C:\> grep /"/"CV5055_active.tab"

Thus making the filename part of the search string and not the actual
target of grep.

==========

OK, Specifics: here is -ONE- record that is causing a problem:

C:\Vanguard\EXTRACTS\CR2>grep "T1202 00TR" CV5055_active.tab
        T1202 00TR  01  "DISP 12CT TRAY 0       EA  [...snip...]

C:\Vanguard\EXTRACTS\CR2>grep "T1202 00TR" CV5055_active.tab | OD -c
0000000       \t   T   1   2   0   2       0   0   T   R           0   1
0000020   \t   "   D   I   S   P       1   2   C   T       T   R   A   Y
0000040   \t   0  \t   E   A  \t      [snipped here]

Note the second line of the dump begins with a tab (\t) followed
immedately by a quote character -- that's the pattern I want to find
elsewhere in the file.

It seems the DOS version of grep doesn't understand "\t" as "tab", which
also makes things tough to find...

> -----Original Message-----
> From: sgvlug-bounces at sgvlug.net 
> [mailto:sgvlug-bounces at sgvlug.net] On Behalf Of BB Odenthal
> Sent: Friday, June 01, 2007 11:50 AM
> To: SGVLUG Discussion List.
> Subject: Re: [SGVLUG] Grep and the command line...
> 
> 
> The extended regex /\t"?/ will search for a tab followed by 
> zero or one double quotes. So this will match the following:
> 
> <Tab>"
> <Tab>
> 
> You can also write it as /\t"{0,1}/
> 
> -bb
>   
> 
> -----Original Message-----
> From: "Emerson, Tom (*IC)" <Tom.Emerson at wbconsultant.com>
> Date: Fri, 1 Jun 2007 11:37:16 
> To:"SGVLUG Discussion List." <sgvlug at sgvlug.net>
> Subject: [SGVLUG] Grep and the command line...
> 
> First off, I'll admit I'm NOT a regex "guru", but I'm getting 
> better. Now I have a search that sorta has me stumped (and 
> part of the problem is that I'm doing this using the 
> microsoft "SFU"(*) version of grep at the DOS command line)
> 
> Here's the deal: I have a tab-delimited file that appears to 
> have a mismatched "quote" somewhere.  For example, the file 
> /should/ look like
> this:
> 
> 
> -->key   <tab>description.....      <tab>123<tab>456<tab>etc.
> 
> Note that "description" is an essentially free-form text 
> field.  Any actual "tabs" in the data (description) have been 
> converted to "<t>". Somewhere in the file I have this:
> 
> -->key   <tab>"description....       <tab>123<tab>456<tab>etc.
> 
> Note the subtle inclusion of a double-quote character at the 
> begninning of the description, but not at the end.  This 
> "quote" character is part of the actual data.  As a result, 
> the process importing the file sees the initial quote 
> character, strips it from the field, and looks for a 
> corresponding close-qoute.  Since it doesn't see it before 
> the CR/LF, it includes THE REST OF THE LINE as part of the 
> description!
> 
> So, how do I search for the sequence <tab><double-quote>?
> 
> Bonus points for "how do I actually type this in at a command 
> line"? (using the aforementioned DOS command line and 
> Microsoft's version of
> grep?)
> 
> Tom
> 
> (*) SFU = Services for Unix -- a freebie from MS that 
> includes a bash shell, tab-completion [within DOS!] and a 
> fairly standard assortment of "basic tools"
>