i'm using crossfilter large set of data dimension grouped timestamp (epoch millis). i'm doing reduce sum on 1 of values , seems correct until last value.
i've prepared simple code:
var group = api.getdimension().group(function (d) { return d.wholepropertypowertimestamp; }); group.reducesum(function(d) { if (d.wholepropertypowertimestamp === 1489428000000) { console.log("d.wholepropertypower: " + d.wholepropertypower); } return d.wholepropertypower; }).top(1);
very simply, takes dimension, applies timestamp group values, runs reduce sum on , if timestamp time on march 13th (this diagnostic, last value , 1 sum incorrect), console output show values before summed.
it complex post meaningful representation fiddler (or analog). best can a screenshot of simple test in dev environment. expected result (a manual sum of console logs) should 1530.25
, sumreduce gives result of 6184153.75
.
this not sum of values, if modify group use 1 group of natural ordering:
var group = api.getdimension().group();
the summed value of every element's wholepropertypower 9699141.59375
.
how group dimension , implement reducesum function?
Comments
Post a Comment