A simple JVM library with utilitarian methods for working with text files of any size, including merge sorting and binary search. The library is based on the Java NIO and Kotlin coroutines.
Apache License 2.0
3
stars
1
forks
source link
Merge sorting: too many file descriptors open for large files #9
Currently merging is performed in parallel, so n-parts means n open channels, which leads to degrade performance or failure (query timeout in LineReader).
Need to introduce a new control parameter (how many file channels could be open simultaneously) and make the merge partially sequential.
related to #7
Currently merging is performed in parallel, so n-parts means n open channels, which leads to degrade performance or failure (query timeout in LineReader). Need to introduce a new control parameter (how many file channels could be open simultaneously) and make the merge partially sequential.