Field order
I've searched the archives and found this question was asked in 2001 but never answered. Does the order of fields in a table make a difference? In Paradox (from whence I come), there was some belief that reliability was increased if memo (text) fields were placed at the end of the table. If field order does make a difference, does the EMS PostgreSQL manager "reorder" function physically rearrange the fields?
Thanks for your time in considering this basic question
On Wed, Aug 17, 2005 at 12:09:12 -0600,
Michael Schmidt <MichaelMSchmidt@msn.com> wrote:
I've searched the archives and found this question was asked in 2001 but never answered. Does the order of fields in a table make a difference? In Paradox (from whence I come), there was some belief that reliability was increased if memo (text) fields were placed at the end of the table. If field order does make a difference, does the EMS PostgreSQL manager "reorder" function physically rearrange the fields?
Thanks for your time in considering this basic question
Some space can be saved by putting the columns with the largest alignments
first.
The columns are not moved around to do the above, but are kept in the
declared order.
On Wed, Aug 17, 2005 at 03:24:58PM -0500, Bruno Wolff III wrote:
On Wed, Aug 17, 2005 at 12:09:12 -0600,
Michael Schmidt <MichaelMSchmidt@msn.com> wrote:I've searched the archives and found this question was asked in 2001 but never answered. Does the order of fields in a table make a difference? In Paradox (from whence I come), there was some belief that reliability was increased if memo (text) fields were placed at the end of the table. If field order does make a difference, does the EMS PostgreSQL manager "reorder" function physically rearrange the fields?
Thanks for your time in considering this basic question
Some space can be saved by putting the columns with the largest alignments
first.
The columns are not moved around to do the above, but are kept in the
declared order.
Ugh, and here I'd been doing it the other way around. I assume
variable-length stuff should always go last, right?
This make sense?
Variable goes last, always
Larger alignment before smaller
NOT NULL before nullable
--
Jim C. Nasby, Sr. Engineering Consultant jnasby@pervasive.com
Pervasive Software http://pervasive.com 512-569-9461
On Wed, Aug 17, 2005 at 17:40:34 -0500,
"Jim C. Nasby" <jnasby@pervasive.com> wrote:
Ugh, and here I'd been doing it the other way around. I assume
variable-length stuff should always go last, right?
I think it depends on the alignment of the type. Certain text, char and varchar
should go at the end. Nullable columns also make a difference since there is
a shortcut used for rows that contain nulls.