11OLSEN

I noticed a slow down of the query with growing table size/number of columns.

Each row in the table is a vector with normalized values. I'm looking for most similar entrys with a query like this (eucl. distance): SELECT ... , ((column1 -in1)²+ (column2 -in2)²+...up to 100 columns ) as 'distance' FROM table ORDER BY 'distance'

What I wonder now is, can I do the same if I calculate the distance as a fixed value for every row when inserting the row. So every row has a distance field with a precalculated distance from the zero point of the mutidimensional space we are in? Later I could just ORDER BY ABS('distance'-input). Are the results the same? 

I noticed a slow down of the query with growing table size/number of columns.
Each row in the table is a vector with normalized values. I'm looking for most similar entrys with a query like this (eucl. distance): SELECT ... , ((column1 -in1)²+ (column2 -in2)²+...up to 100 columns ) as 'distance' FROM table ORDER BY 'distance'

What I wonder now is, can I do the same if I calculate the distance as a fixed value for every row when inserting the row. So every row has a distance field with a precalculated distance from the zero point of the mutidimensional space we are in? Later I could just ORDER BY ABS('distance'-input). Are the results the same? 




I noticed a slow down of the query with growing table size/number of columns.
Each row in the table is a vector with normalized values. I'm 

optimize sqlite distance query