Archives of the TeradataForum

Message Posted: Tue, 09 Sep 2003 @ 10:38:46 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

I'm probably missing something pretty obvious here, but I can't see the wood for the trees so I'll put this query to the good people on the TDATA list. (This probably reads like one of those irritating impossible problems, but read on.....).

Take two tables, let's call them TABLE_A and TABLE_B, both present within the same database, on the same Teradata server (a NCR5200, 6 node, 60amp, V2R5.0.1.6 running DBS 5.0.1.12, but not that relevant really).

Both contain exactly the same number of rows (69,588,081).

DDL for both is exactly the same (inc. compression, indexes, fallback, etc - after image journals ARE enabled on both to the same target journal table, but I reckon this should be irrelevant). Stats are current on each, are identical, and on the same index & columns.

Both tables have been created via ARCMAIN dictionary copy.

However, DBC.TABLESIZE reports a different amount of total currentperm for each.

TABLE_A is reported as 3,800,815,104 bytes.
TABLE_B is reported as 3,770,984,448 bytes.

Why the 30Mb difference??

I've noticed that the discrepancy is greater on larger tables (although not proportionally), and can be zero on small tables.

Confused??? You will be......

To make matters more confusing, I've tried :-

i) copying TABLE_A to TABLE_A_COPY, with the result that TABLE_A_COPY is smaller (by approx.30Mb)

ii) copying TABLE_B to TABLE_B_COPY, with the result that TABLE_B_COPY is smaller (by approx.30Kb ...table header size, maybe?).

.....however, TABLE_A_COPY and TABLE_B_COPY are now identical in size !!! Which sort of implies, that TABLE_A and TABLE_B should be identical in size as well, or are being represented incorrectly in the first place.

Am I just having a bad day and missing something fundamental, or is this Teradata possessed????

Cheers,

Mark

Mark Dodgson
AM Common Services - Service Development GDW

	<Prev	Next>		<<First	<Prev	Next>	Last>>