Copy to list 2009/1/30 Nigel Thomas <nigel.cl.thomas@xxxxxxxxxxxxxx> > A table with no rows is still a table, so it has to be stored somewhere. > > One or more tables in a heap cluster would not have their own segments (eg > SYS.USER$ is stored in cluster C_USER# along with SYS.TSQ$). Unless the > cluster name is the same as (one of) the table(s). > > A partitioned table has no segment (the partitions each have a segment). Of > course one of the partition names may be the same as the table name (eg see > SYS.LOGMNR_IND$ in the XE install). > > Regards Nigel > 2009/1/30 <Joel.Patterson@xxxxxxxxxxx> > > >> I would like to clarify or confirm. I can query (user_tables union >> user_indexes) minus user_segments and see about 4 tables and 9 indexes. >> 7 turn out to be normal and 2 are IOT - TOP indexes. >> >> I guess the 3 tables with 0 rows and all the indexes that go with them >> might not have a segment? >> >> There is one table with and IOT - TOP index with 100000 rows. Two >> indexes go with this table, one of them is IOT - TOP, the other normal. >> >> Even and Index has a segment, why would I not have a segment for >> this(ese) table(s) or Indexes? >> >> >> >> >> -- >> //www.freelists.org/webpage/oracle-l >> >> >> >