[NLP-er]

Originally Posted by Dave Hybrid

I don't know what you did but it's slow, really slow, unbearable. Something isnt right with your changes.

Of course - you just wrote that you didn't made UNIQUE INDEX - so you just came back to time which you had in your very first release.

Once again:
Delete data duplication like described here (note this is very important - otherwise unique indexes will not be set)

And you had data duplication:
#1062 - Duplicate entry

So you don't have index now - it have to be slow

My changes are ok and make it works faster - just have to make all steps.

If you had error with data duplication - drop duplicated data. If you did it and have same problem - drop data duplication again - it just means that in the meantime between dropping data duplication and trying to create UNIQUE INDEX, some new data came with duplications (google doesn't sleep).

Before you will drop duplicated data again it is wise to set normal index again:

create INDEX originaltext on wt_cache_short (originaltext, tl);

then drop data duplication and reindex to unique as described in previous post

As I wrote - you always have the easy way. You want keep your cache - made everything as described and don't complain that with my update is something wrong if you didn't make it.