a suite of four utilities to scan a set of files looking for a set of strings:
extract  - displays all lines in all files with a match on one or more strings, case-sensitive.    (precise)
extracti - displays all lines in all files with a match on one or more strings, case-insensitive.  (relaxed)
without   - displays all lines in all files with no match on any of the strings, case-sensitive.   (precise)
withouti  - displays all lines in all files with no match on any of the strings, case-insensitive. (relaxed)
It works on text and html files.
To find matching lines in a single file, type:
  java.exe -jar C:\com\mindprod\extract\extract.jar "wombat" "zebra" - myfile.html
Everything is case-sensitive. The results display on the
console. You can redirect with in the usual way with >
results.txt
You can also list several files on the command line:
  java.exe -jar C:\com\mindprod\extract\extract.jar "wombat" "zebra" - myfile.html C:\mydir\another.html
To use search the current directory of files:
  java.exe -jar C:\com\mindprod\extract\extract.jar "wombat" "zebra" - .
Sorry no wildcards, just . , and ..
DON'T USE WILDCARDS (*.xxx) unless you deeply understand how
they work. See http://mindprod.com/jgloss/wildcard.html.
Windows expands them, not the utility, and feeds them to the
utility as a giant list of all the directories and files in
the current directory. The utility will thus tend process
all the files in your directories, when you just meant to
process the files in the current directory.
The -s switch makes all subsequent directories searched
recursively to include all their subdirectories.
e.g.
  java.exe -jar C:\com\mindprod\amper\extract\extract.jar "wombat" "zebra" - -s E:\mindprod
will scan all files in the mindprod directory tree.
INVOKING
If you have the jar extension set up as executable, you can
abbreviate:
  C:\com\mindprod\extract\extract.jar "wombat" "zebra" - E:\mindprod
If you have Jet, you can compile the jars and abbreviate even further:
  extract.exe "wombat" "zebra" - E:\mindprod
Instead of extract, you can use one of the other utilities
that work the same way: extracti, without, withouti.
EXTENSIONS
Because extract is designed to work only with text file, it ignores all files
except those with the following extensions:
ans, asm, bat, batfrag, btm, btmfrag, c, cfrag, cmd, cpp,
cppfrag, css, cssfrag, csv, csvfrag, ctl, doc, dtd, dtdfrag,
e, h, hfrag, hpp, hppfrag, htm, html, htmlfrag, ih, ini,
java, javafrag, jnlp, jnlpfrag, jsp, jspfrag, list, log, look,
lst, mac, mft, pas, policy, prn, properties, ps, rh, sh,
site, sql, sqlfrag, tab, txt, use, wiki, xml, xmlfrag, xsd,
xsdfrag.
If you need more extensions, please ask, or add them yourself in Extract.java.
SWITCHES
-all
You may use the -all switch on the command line to extract
only if all the strings match.
-where
You may use the -where switch on the command line if you
want to have the output include the name of the file and
line number where the line was found in CSV format.  It
appears on the console. If you want to capture it to a file
use > redirection.
-s
-s means include files in all subdirectories of a directory
mentioned.
-
Dash separates the strings and regexes from the file names.
QUOTING
Depending on your operating system, there are a number of
characters that have magic meaning on the command line. They
won't necessarily be passed through to the program. Ones to
watch out for in Vista include " \ & ^ + | < > space.
Try enclosing awkward characters in quotes e.g. "<& | >"
To put a " inside quotes use \" e.g. "he said \"Hi\"".
To put a \ inside quotes use \\ e.g.
"the \\ is called backslash; the / is called slash".
If these drive you to distraction, you might try using
regexes, since the OS does not interfere with them in any
way.
REGEXES
Regexes are an advanced feature primarily for programmers.
The regexs to search for are put one per line in a file
with any extension.
Read up on how they work at http://mindprod.com/jgloss/regex.html
java.exe -jar extract.jar "apple" "pear" @myregex.txt - somefile1.txt -s somedir
extract.jar "apple" "pear" @myregex.txt - somefile1.txt -s somedir
extract.exe "apple" "pear" @myregex.txt - somefile1.txt -s somedir
Because command shells interfere so much with special
characters on the command line, it would be highly confusing
to put regex strings directly on the command line.  Instead
you put your regexes in tiny separate files.  You but them
on the command line to the left of the dash, with a leading
@, not part of the filename itself.
The tiny files must be incoded in UTF-8.  They don't include
Java string quoting.  They don't use \uxxxx the way Java
strings to you just key your accented characters directly.
Use the Quoter amanuensis
http://mindprod.com/applet/quoter.html to help you compose
your regex strings with just regex quoting, not Java string
quoting.
For example if you wanted to search for either \ or / just
before the word html, HTML hTMl... (case-insensitive) you
would code (?i)[\\/]html in your little regex file not
"(?i)[\\\\/]html" as you would in a Java program.
It does not matter whether you use extract, extracti,
without, withouti, the regexes are all case-sensitive unless
you use (?i) embedded in the regex to turn on case
insensitive mode.  You can go back and forth in the same
regex using (?-i) to flip back to case-sensitive mode.
You can  mix and match as many regular search strings and
regexes as you can fit on the command line. You can speed
the program up slightly by putting the most likely matches
first. The results are the same no matter what order you put
them.
Be careful that your editor does not trim or add any
trailing blanks on your regexes. You are safer to use
\p{Blank} or \s (which has slightly different meaning).
Any control characters in your regex files are stripped out
prior to use. This means you can break your regexes into
several lines in the file without penalty.
LIMITATIONS
1. Extract just does searches, no replacements.
2. Extract does not show you to context of where it found
each line, just the line itself.
3. Extract is a batch, command line program. There is no GUI
and no interaction.
4. Extract offers no debugging tools to help you figure out
why your regexes failing to find the lines you expect it to.
5. Regexes cannot span lines.
6. Java-style regexes only, no Unix, Perl, Funduc, SlickEdit etc.
Why the haystack logo?  These utilities help you find your
needles in a haystack.