Strange situation here,
We noticed that if we ask a package to return the next value which would be
used for a table entry (done using a sequence), we get a number.
If we go through a local synonym to ask the same package to answer the same
question, we receive a slightly different answer.
Is it possible that there are two caches of sequences values, depending on
how the package which refers to the sequence is executed? I suppose if two
users are entering records concurrently in the same table, they each have
their own set of cached sequence values (?).
We don't think this happened in 11g, it only seems to be happening in our
I will log an SR as well, just thought I might ask this list first.
My profiles: [image: Facebook]