Hi Elmer

Yep! That's what I've seen. Test was opened by the SQL engine, and the SQL engine closed it since it opened it. Just like when you ...

SELECT field1 from test a1 into cursor c_test

. There is no "a1" alias left open.

>Isn't that due to the fact that VFP SQL operates on the table on the disk and not the aliased workarea by opening the underlying table when you specify the alias name in the "select from " as in
>
>

>select * from (dbf("test_in")) into cursor test
>

>
>I would think that it is using the table name as the alias for the second instance.
>
>However, what I find as weird is that after executing the following 3 series of commands, the data session does not show "TEST" in the list of open workareas if the underlying table was already open in an aliased workarea of a different name.
>

>close data all
>use test alias test_in
>select * from test_in into cursor test1
>*!* the data session contains "TEST_IN" and "TEST1" as open workarea aliases
>*!* but not "TEST"
>
>close data all
>use test alias test_in
>select * from test into cursor test1
>*!* the data session contains "TEST_IN" and "TEST1" as open workarea aliases
>*!* but not "TEST"
>
>close data all
>select * from test into cursor test1
>*!* the data session contains "TEST" and "TEST1" as open workarea aliases
>

>With Glenn's error and the above examples, that seems to suggest to me that VFP opens another instance of test.dbf for each of the select statements, but automatically closes "TEST" workarea if the underlying table was already in use in another workarea. Can that be?
>
>
>>BTW, I get the same error in VFP7.
>>
>>>I get this error in the following:
>>>
>>>

>>>create table test ( ctest c(10))
>>>append blank
>>>append blank
>>>append blank
>>>use
>>>
>>>use test alias test_in
>>>select * from test_in into cursor test
>>>

>>>
>>>The alias test is NOT in use !
>>>
>>>Glenn