hi,
if you do an import, is it better to set the statistics option on the import itself, or should we gather_schema_stats on the schema after the import, or is it not necessary at all?
All I am trying to do is refresh the database on the target machine with the data from the source machine, using export full=y and then importing fromuser for just the one user.
I thought (possibly wrongly) that you should gather new statistics based on how the data looks now after the import.
Are you saying that if I let the import do the stats that they will basically be exactly the same on the new instance as they were on the old one - which in itself relies upon how often stats have been gathered on the source machine in the first place I presume, and therefore they could be rubbish if not done lately?
If this is acceptable, then I will do it that way.
I was trying to find out the most efficient way to get stats associated with the data on the new instance, so that the optimizer will have the latest set of stats to go on, so was thinking along the lines of doing the gather stats afterwards.
Presumably if I did no stats at all then performance would be very slow as the optimizer would hav nothing to go on to build execution plans?
Bookmarks