From: James on
MitchellWmA wrote:
> After intensive searching found an app to almost do what I need. I'm
> mentioning it only to give an example.
>
> $ware Super Text Search v2.8 will search text files in a folder for
> text strings
> http://www.galcott.com/textsrch.gif
> http://www.digitalriver.com/dr/v2/ec_MAIN.Entry10?xid=22609&PN=1&SP=10023&V1=187027&CID=137331&CUR=124&DSP=&PGRP=0&ABCODE=&CACHE_ID=0
> but it can't find all words beginning with a letter and ending with
> another letter.
>
> i.e., the best it can seem to do is this - when I tested it by asking
> for "s*e" to see if it could find "software" or "smile" or safe" (in
> its own text string search format, of course), it only could return
> things like "she", "see" and "sue". No wildcard that I could find
> could pull up words with more than 1 letter in between the start and
> end.
>
> So I'm stumped. If anyone knows of something that can do this please
> let me know.
>
> The text string search programs I've tried don't do this. Or if they
> can, they don't display the text easily in a list. They just pull up
> a list of files that you have to open and then hunt around for the
> word(s).
>
> After 2 days looking and looking every spare moment I'm bugging the
> group again. tia :)

Throw the text files you want to search through on a Linux box and do a
"grep <string> *".


James http://cronus.ath.cx
From: James on
MitchellWMA wrote:
> On Mar 29, 7:54 am, j...(a)webace.com.au wrote:
>> On Mar 29, 8:19 pm,MitchellWmA<nospamm...(a)nonsense.com> wrote:
>>
>>> but it can't find all words beginning with a letter and ending with
>>> another letter.
>> Why not do a search with the 1st & last letter missing.
>
> Okay, you lost me *lol*. There would _be_ no search without the first
> and last letter.
>
> Guess I didn't explain well.
>
> In any given puzzle, in this case specifically, a syllacrostic, I'll
> have the beginning letter and the last letter as a definite. What I
> must do is find words that begin and end with those letters to be able
> to build the puzzle. Though I didn't mention it because I didn't want
> to confuse the issue, the word lists involved are very large. I
> currently have about 10 of them, one which has over 169,000 words to
> give you an example. I need to find the words fitting the above
> criteria. Without the first and last letter, I'm left with nothing to
> search for so it's kind of difficult to see what you mean *lol*. I
> expect it's a case of not having thought it through on your part.
>
> Also a simple windows search doesn't cut it either for the very same
> reason. There are going to be words on those 10 word lists that fit
> the bill so I should get several words coming out of each for each
> search. So a windows search is silly because it doesn't tell me
> anything and no words are listed so I'm right back to square one. I
> need something like Super Text Search that will list the actual words
> in an easily viewable format. It does that. What it doesn't do,
> _it_seems_, is find the longer words. I'll illustrate s*e again. A
> text string wilcard search starting with 's' and ending with 'e'
> should bring up a variety of words like this:
>
> software
> smile
> she
> science
> safe
> snake
> spike
> secrete
> systematize
> sacrifice
> sponge
>
> They all begin with an 's' and end in an 'e' but are all very
> different. Leaving off the first and last letter wouldn't help here.
>
> Anyway I'll keep searching for a freeware for this of course. It's
> just that if someone knows of something already it'd be nice to know
> of it. It might happen like the text stats where there wasn't a good
> solution to that software need but I won't know if I don't ask.
>
> Surprising though that a $ware program doesn't seem to do this.
> Everything I tried, at any rate didn't bring up these words only
> ones that had one letter in between. The help file didn't contain
> anything about wildcard searches either. Not so Super a text search
> $ware, if you ask me *g*.
>

Sorry, now that I read the full description of the problem...you would
use "grep -E s[a-z]+e *"

I created a text file given the word list you provided above and this
command returned all values. The result of following the directions in
my first post only returned 2 of the 11 results.


James http://cronus.ath.cx
From: »Q« on
In <news:iKSdndBFUtaP55DbnZ2dnUVZ_u-unZ2d(a)comcast.com>,
James <deadjimmy(a)comcast.net> wrote:

> Sorry, now that I read the full description of the problem...you
> would use "grep -E s[a-z]+e *"
>
> I created a text file given the word list you provided above and this
> command returned all values. The result of following the directions
> in my first post only returned 2 of the 11 results.

You'll get a lot of false positives (fragments of words) with that
expression:

$ echo "astute researchers see serpents asleep" | grep -Eo s[a-z]+e
stute
searche
see
serpe
slee

Fix with the -w option:

$ echo "astute researchers see serpents asleep" | grep -Ewo s[a-z]+e
see

(There are other ways, but that seems simplest to me.)

You'll also miss one very short word, se (as in "per se"). Fix by
using * instead of + in the expression.

--
�Q�
From: James on
�Q� wrote:
> In <news:iKSdndBFUtaP55DbnZ2dnUVZ_u-unZ2d(a)comcast.com>,
> James <deadjimmy(a)comcast.net> wrote:
>
>> Sorry, now that I read the full description of the problem...you
>> would use "grep -E s[a-z]+e *"
>>
>> I created a text file given the word list you provided above and this
>> command returned all values. The result of following the directions
>> in my first post only returned 2 of the 11 results.
>
> You'll get a lot of false positives (fragments of words) with that
> expression:
>
> $ echo "astute researchers see serpents asleep" | grep -Eo s[a-z]+e
> stute
> searche
> see
> serpe
> slee
>
> Fix with the -w option:
>
> $ echo "astute researchers see serpents asleep" | grep -Ewo s[a-z]+e
> see
>
> (There are other ways, but that seems simplest to me.)
>
> You'll also miss one very short word, se (as in "per se"). Fix by
> using * instead of + in the expression.
>

Excellent! =D I hadn't thought of this as the examples provided were
all longer than 2 characters, and seemed to fit some basic criteria (all
words beginning on a new line, and all began with S and ended with E).
I hadn't the foresight to think of what the results might be should the
entire list not follow this criteria.

Thanks Q, "you done learn-ed me somethin!"


James http://cronus.ath.cx
From: MitchellWmA on
On Fri, 30 Mar 2007 00:49:05 GMT, jon
<mcdaddums3(a)hhotmail.ccom.invalid> wrote:

>
>MitchellWMA wrote:
>
>>^s[a-z]*e$
>
>I should have read the group a bit more before posting my last post.
>:-)
>
>>So my money is on Agent Ransack.
>
>Well done. I've seen it referenced here so often but have yet to look
>at it.
>Now, I really must.

I even already had it here at home. I just didn't know about "regular
expressions" and all that as it was all new to me. So I'll probably
really be learning how to use Agent Ransack now that I'll be building
puzzles. It will just happen because Ill need to learn how to
optimally and successfully "trawl" through the vast number of words in
all my separate word lists.

Again, I have to really thank the group. You can't look for something
if you don't know what it is ... now that I know about this regex
business, any new word search needs I'll be able to google for and
then use AR! And the price is the best ... :)
First  |  Prev  |  Next  |  Last
Pages: 1 2 3 4 5 6 7 8
Prev: Freeware needlepoint design software?
Next: protect..