I am not aware of being able to go backwards when reading a file (other than using the seek()
method) but I might be wrong...
A simpler approach would be to:
Create a dictionary, key value being the long numeric values while the value being the line to which it belongs:
<2265507967,07 IVIN 15:37 06/03 022 00:00:14 600 2265507967 0:03>
Go through the file one line at a time and:
a. If the line contains
ITRS
, get the value from the line and check your dictionary. Once you will have found it, clear the dictionary and go back to step 1.b. If it does not contain
ITRS
, simply add the number and the line as key-value pairs.
This should be quicker than going through one line at a time and also simpler. The drawback would be that it could be quite memory intensive.
EDIT: I do not have a .NET compiler handy, so I will provide some pseudo code to better explain my answer:
//Initialization
Dictionary<string, string> previousLines = new Dictionary<string, string>();
TextReader tw = new TextReader(filePath);
string line = String.Empty;
//Read the file one line at a time
while((line = tw.ReadLine()) != null)
{
if(line.contains("ITRS")
{
//Get the number you will use for searching
string number = line.split(new char[]{' '})[4];
//Use the dictionary to read a line you have previously read.
string line = previousLines[number];
previousLines.Clear(); //Remove the elements so that they do not interrupt the next searches. I am assuming that you want to search between lines which are found between ITRS tags. If you do not want this, simply omit this line.
... //Do your logic here.
}
else
{
string number = line.split(new char[]{' '})[4];
previousLines.Add(number, line);
}
}