Can such operation be tuned?
I have an inverted index stored in a RandomAccessFile. Whenevr a search is taking place the procedure is as follows:
1) read seek-point into the file
2) read a list of numbers
3) one of the above numbers points to the next list of numbers somewhere else in this file.
4) seek into this next point in the file and continue the same cycle until you reach start of file.
Now, if all these lists where consecutive one could read 1-2MB of data and do some sort of buffering. But if the data is spread across the RandomAccessFile, is there a way to optimize the read access?
NOTE: The file is to big to be loaded as a whole
Any ideas are very appreciated.