I forgot to output the original phrase column...

I forgot to output the original phrase column...

The trick is to use a column-value sub-query which selects a list of the words in the phrase which do not occur in the list of noise words and then use FOR XML PATH('') to put the split words back together into a single value.

The trick is to use a column-value sub-query which selects a list of the words in the phrase which do not occur in the list of noise words and then use FOR XML PATH('') to put the split words back together into a single value.

For my testing I found converting the target column to DATE for time insensitive comparison performed over 55k rows actually performed better than using full DateTime with filter values inclusive of the time needed, and I find both to be accurate. I di...

For my testing I found converting the target column to DATE for time insensitive comparison performed over 55k rows actually performed better than using full DateTime with filter values inclusive of the time needed, and I find both to be accurate. I di...

A simple way to think about it is; FROM defines the data that is accessible, and then WHERE restricts the data accessible by FROM, and then GROUP BY groups the data restricted by WHERE, and then HAVING restricts the data grouped by GROUP BY...

Another solution that does not use partitioning; SELECT DISTINCT --- convert all non-intersecting left-side values to NULL CASE a.attrib WHEN b.attrib THEN a.attrib ELSE NULL END 'left', b.attrib 'right' FROM @demo2 a LEFT OUTER JOIN @demo b...

Another solution that does not use partitioning; SELECT DISTINCT --- convert all non-intersecting left-side values to NULL CASE a.attrib WHEN b.attrib THEN a.attrib ELSE NULL END 'left', b.attrib 'right' FROM @demo2 a LEFT OUTER JOIN @demo b...