Like a bloom filter, a bitmap index represents membership. A bitmap index is an array of bits, a bit represents each row where each bit set to 1 represents membership in the set.
In addition to the bitmap indexes, we need a map that maps index positions to rows in the database.
Bitmap indexes do well at indexing categorial data with well defined values like countries, dates and ages as a few examples. Bitmap indexes don’t do well with numerical data with continuous values
A 1 in the array of bits represents membership in the set for a unique term which makes them suitable for WHERE A = B type evaluations and not WHERE A > B evaluations.
Using bitmap indexes means managing a lot of indexes because you need a new array of bits per unique term.
Read full article from Using bitmap indexes in databases « kellabyte
In addition to the bitmap indexes, we need a map that maps index positions to rows in the database.
Bitmap indexes do well at indexing categorial data with well defined values like countries, dates and ages as a few examples. Bitmap indexes don’t do well with numerical data with continuous values
A 1 in the array of bits represents membership in the set for a unique term which makes them suitable for WHERE A = B type evaluations and not WHERE A > B evaluations.
Using bitmap indexes means managing a lot of indexes because you need a new array of bits per unique term.
Read full article from Using bitmap indexes in databases « kellabyte
No comments:
Post a Comment