I have a txt file with some data that looks like this:
a:1(2,3) 55(33,45,67)
b:2(1,33,456) 4(123,12444)
which means that word "a" appear in text 1 in places 2 and 3 and in text 55 in places 33,45 and 67..
I have some texts and I go all over those texts and if I see that the word "a" appears in a text then I need to update the text file above accordingly. (the data about "a" can be big开发者_Go百科ger than a line of course)
How can I update the line where "a" is without damaging the line where "b" is? I saw here in stackoverflow that I can use maybe xml file. If I use xml file can I read the whole "a" data, update it and then write it again without damaging the "b" line? Or maybe each data about a word can be in some data structure that I can somehow read from the file, update it and then write it to the same position?
Thanks in advance,
Greg
You might encode the same information in XML with
<words>
<word name="a">
<text id="1">
<place id="2" />
<place id="3" />
</text>
...
</word>
...
</words>
As you can see, XML has the same issue as your text format: updates require rewrites, so it won't be as much help as you're hoping.
Given the sample in your question, I assume you're not indexing a huge corpus. If so, perform updates in the following steps:
- Read the current index file into the data structure you're using for the index.
- Update your data structure to reflect the contents of the newly added texts in the corpus.
- Write the contents of the entire updated index to a new temporary file.
- On success, rename the temporary file to the master index file from step 1.
If you do have a very large corpus, the text-based index is inappropriate. Consider using a real database.
If you change this text file to XML, you can easily manipulate the file using LINQ to XML.
Take a look here. Specifically, the manipulation section.
using: string.Insert
but i would use a 2D array int Places[text][place].
your question has nothing to do with XML though you can use XML for this, a simple data structure, would be simpler....
edit:
ok use this:
dictionary<string,dictionary<int,list<int>>> .
use a dictionery the key would be your word the value would be another dictionery with the text numbers and the locations... you can serlize all of this to file by the way....
精彩评论