China DOS Union

What I need to do is like this:
There is an XML message, about 1 MB in size, and the longest line is a little over 4000 bytes. In that case, the find command can't search for keywords in such ultra-long lines.
Current ideas:
1, use the findstr command; I've tested it, and it can search 4 KB lines
2, split the long lines into short lines; this is basically a matter of parsing the message.
I hope friends with similar experience can discuss it together.

[ Last edited by amark on 2009-2-9 at 11:20 ]

. Paste a sample here so we can have a look

xmlstarlet
F http://baiy.cn/utils/f/index.htm

SED
AWK

and so on, tools like these should be up to the job

Since the findstr command can search 4 KB lines, why not use it?

Uh, here's the situation.
After findstr searches and returns that 4 KB line, I still need to find specific key fields within those 4 KB.
For example:
<errorfields><field name="date" type="0" value="" value1=今天" value2="昨天" time1="20090204" time2="20090203"/></errorfields>
and extract several values from this section, such as "今天" "昨天" "20090204" "20090203".
Yesterday I tried using for /f "tokens=31* delims=<> usebackq" %%i in (error.txt) to do this step.
As a result, %%i could only get the content of the 31st node, such as attribute name="television" type="0" value="1"
But when I wanted to further process %%j, that is, the string containing the required key fields, since the xml is UTF-8 encoded, echo %%j would produce a pile of garbled text.
I'm sorting out my thinking right now.
PS: due to limitations, I can't use third-party tools. Thanks in advance for the suggestion in reply #2

[ Last edited by amark on 2009-2-4 at 09:11 ]

Then convert it to ASCII and use third-party processing,
I think for /f may not be able to handle long fields over 4k, and with many nodes it definitely can't handle it.
Unless you use some other way to split the lines again.

At most the code would be a bit more complicated; batch processing should be able to cut out the required strings from XML.
Upload a file and we'll know at a glance

I finished it last week. I called one vbs to replace > with >/n, which solved the long-line problem. I called another vbs to convert utf8 to unicode, then type it to ansi. That solved the garbled text problem. Thanks here to all the posters above for their opinions and suggestions.

Post the code too, let everyone share it