mrtomrod / binaryfilesearch Goto Github PK

What do you think about moving the code for _enter_() to the end of _init_()? I don't always want to indent my entire program, but rather just store the open file in a variable. As written, I need to call _enter_() explicitly, which feels a bit strange.

Super library by the way. Just what I was looking for.

The sort suggested by the current README does not make BinaryFileSearch work on keys (first column) of varying length.

I figured out this is because the module expects strict prefixes to come first, while bash sort puts them at the end. For example, the module expects

ban
banana
cat

whereas sort produces

banana
ban
cat

We can obtain the expected sort by first padding the keys with whitespaces (assuming the keys don't have whitespaces or characters strictly before in ASCII) in awk, and removing them after sorting:

awk 'BEGIN {FS=OFS=","} {s = sprintf("%*s",N-length($1),""); print $1 s,$2}' input.csv \
 | LC_ALL=C sort --key=1 --field-separator=','
 | sed 's/ //g' > output.csv

where N is the max size of keys, which can be obtained using:

cut -d ',' -f 1 input.csv | wc -L