Ocke Janssen wrote:
Moin Drew,
Drew Jensen wrote:
| AVG
| MAX
| MIN
| SUM
| EVERY
| ANY
| SOME
| COUNT
| STDDEV_POP
| STDDEV_SAMP
| VAR_SAMP
| VAR_POP
| COLLECT
| FUSION
| INTERSECTION
Like Andrew earlier I'm wondering if I am just being tired here..but
I give up...
I found it in http://www.wiscorp.com/sql200n.zip
6WD2_02_Foundation_2007-12.pdf page 555 and thought when supporting
SQL than I support a little bit more while touching the code. As far
as I know hsqldb doesn't supports this at the moment.
What do fusion and collect do?
You'll find an explanation on page 561
ix) If FUSION is specified, then the result is the multiset M such
that for each value V in the element
type of DT, including the null value, the number of elements of M that
are identical to V is the
sum of the number of identical copies of V in the multisets that are
the values of the column in
each row of TXA.
x) If INTERSECTION is specified, then the result is a multiset M such
that for each value V in the
element type of DT, including the null value, the number of duplicates
of V in M is the minimum
of the number of duplicates of V in the multisets that are the values
of the column in each row
of TXA.
Well OK "my bad", wouldn't it be reasonable to do this as a statistical
addon java library distributed as an extension maybe.
Which might be what you're getting at with -
PS: May be we should not offer fusion, collect and intersection in the
UI.
then again adding three new aggregate functions is not that big a time
hit really (you would know better then anyone what that hit would be) -
and - what is the point of working on a FOSS package if you don't get to
scratch your own itch every now and then. I sincerely mean that.
- but...if you are really looking for some enhancements for the parser,
LOL, I'll be bold and propose that any of the following would have a
much higher ROI for your time in regards to the number of users that
would find it useful, IMO:
Query Defs that support
UPDATE / INSERT / DELETE / SELECT INTO
with EscapeProcessing = TRUE
OR
Ability to reference aliases within the select statement. i.e.
SELECT *, DATEDIFF( 'dd', DATE_COL, CURRENT_DATE) AS AGING
FROM T1
WHERE AGING >= 30 AND AGING <= 59
OR
UNION | INTERSECT | EXCEPT with EscapeProcessing = TRUE
AND
If a common named parameters is used in multiple select statements
therein, then prompt only once and use it in each.
OR
Ability to create
SELECT DISTINCT A.A1, AB.B1, BC.C1, C.C2
FROM
A LEFT OUTER JOIN AB ON A.A1 = AB.A1
LEFT OUTER JOIN BC ON AB.B1 = BC.B1
LEFT OUTER JOIN C ON BC.C1 = C.C1
using the GUI designer
Just some thoughts.
Take them for what ever they are worth.
Respectfully,
Drew
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
