Hi

We have a table with a LOB column (varchar(MAX)) which we update to NULL, after a period of time.

Historical date sensistive archive / operational data... used for tracking and processing.

Row depth is millions; width is 15 columns (only 1 LOB).

DBCC SHOWCONTIG scanning 'EmailQueue' table...

Table: 'EmailQueue' (244195920); index ID: 1, database ID: 32

TABLE level scan performed.

- Pages Scanned................................: 3485182

- Extents Scanned..............................: 437322

- Extent Switches..............................: 438993

- Avg. Pages per Extent........................: 8.0

- Scan Density [Best Count:Actual Count].......: 99.24% [435648:438994]

- Logical Scan Fragmentation ..................: 0.44%

- Extent Scan Fragmentation ...................: 2.23%

- Avg. Bytes Free per Page.....................: 7347.2

- Avg. Page Density (full).....................: 9.23%

Tables is huge, but empty...

5 of the fields are defined as varchar(4000) and as far as I understand, only consume space based on the actual content. Is this impacting the free space? [These are not impacted by the nulling process].

Essentially, the bulk of the data allocated is allocated to the field that is nulled, but the free space is not allocated to subsequent inserts. It just grows and grows... ideas?