Re: Row pattern recognition

Hi Henson,

I found following in rpr_base.sql:

-- {0} is not allowed (min must be >= 1)
SELECT id, val, COUNT(*) OVER w as cnt
FROM rpr_quant
WINDOW w AS (
ORDER BY id
ROWS BETWEEN CURRENT ROW AND UNBOUNDED FOLLOWING
PATTERN (A{0} B)
DEFINE A AS val > 1000, B AS val > 0
)
ORDER BY id;

However in my uderstanding the SQL standard allows A{0}.

PATTERN (A{0} B)
is equivalant to:
PATTERN (B)

Another interesting PATTERN is:

A{,}

This is equivalant to A{0,}

BTW, after studied this more, I found that A{0,0} is not allowed. In
this form the right hand side number shall be greater than 0. From
ISO/IEC 9075-2 7.9 <row pattern syntax> "Syntax Rules 20)

"If <left brace> <unsigned integer> <comma> <unsigned integer> <right
brace> is specified, then let VUI1 and VUI2 be the values of the first
and second <unsigned integer>'s, respectively. VUI1 shall be less than
or equal to VUI2, and VUI2 shall be greater than 0 (zero)."

However according to the Google, Oracle and Snowflake allows A{0,0}:
they break the standard. So, what do you think PostgreSQL should do
here? My preference is "always follow the standard". But others might
think differently.

Best regards,
--
Tatsuo Ishii
SRA OSS K.K.
English: http://www.sraoss.co.jp/index_en/
Japanese:http://www.sraoss.co.jp