我将以此为开头说我发现的所有文档都不适用于我的oracle版本: https://docs.oracle.com/cd/E51711_01/DR/WeekDay.html; https://docs.oracle.com/cd/E37483_01/server.751/es_eql/src/ceql_functions_date_extract.html
或我可以使用的功能参考没有提及星期几: https://docs.oracle.com/cd/B28359_01/server.111/b28286/functions052.htm#SQLRF00639
,因此我只想找到使用to_char进行此操作的方法,然后进行varchar2比较。如果可能的话,我只想坚持日期格式。
到目前为止,这是我的SQL
select dateSold from sales
where extract(dateSold, day_of_week) in (1, 3, 4, 7)
SQL Error: ORA-00904: "DAY_OF_WEEK": invalid identifier
答案 0 :(得分:4)
您需要使用to_char()
来获取日期,但是您可以将其转换为数字:
select dateSold, sum(quantity) from sales
where to_number(to_char(dateSold, 'D')) in (1, 3, 4, 7)
但是D
是NLS依赖的,因此,如果在美国和法国进行的会话中运行此命令,则会得到不同的结果。这也许就是您看到字符串比较的原因,因为您至少可以控制更多:
select dateSold, sum(quantity) from sales
where to_char(dateSold, 'DY', 'NLS_DATE_LANGUAGE=ENGLISH') in ('MON', 'WED', 'THU', 'SUN')
要查看NLS设置的不同之处,将显示D
和DY
值,用于相同的数据和查询,就像在美国运行一样:
alter session set nls_territory = 'AMERICA';
alter session set nls_language = 'ENGLISH';
with cte (dateSold) as (
select date '2018-08-01' + level - 1 from dual connect by level <= 7
)
select dateSold,
to_number(to_char(dateSold, 'D')) as d,
to_char(dateSold, 'DY') as dy,
to_char(dateSold, 'DY', 'NLS_DATE_LANGUAGE=ENGLISH') as dy_english
from cte;
DATESOLD D DY DY_ENGLISH
--------- ---------- ------------ ------------
01-AUG-18 4 WED WED
02-AUG-18 5 THU THU
03-AUG-18 6 FRI FRI
04-AUG-18 7 SAT SAT
05-AUG-18 1 SUN SUN
06-AUG-18 2 MON MON
07-AUG-18 3 TUE TUE
然后好像在法国运行一样:
alter session set nls_territory = 'FRANCE';
alter session set nls_language = 'FRENCH';
with cte (dateSold) as (
select date '2018-08-01' + level - 1 from dual connect by level <= 7
)
select dateSold,
to_number(to_char(dateSold, 'D')) as d,
to_char(dateSold, 'DY') as dy,
to_char(dateSold, 'DY', 'NLS_DATE_LANGUAGE=ENGLISH') as dy_english
from cte;
and the same thing as if run in the USA:
DATESOLD D DY DY_ENGLISH
-------- ---------- ---------------- ------------
01/08/18 3 MER. WED
02/08/18 4 JEU. THU
03/08/18 5 VEN. FRI
04/08/18 6 SAM. SAT
05/08/18 7 DIM. SUN
06/08/18 1 LUN. MON
07/08/18 2 MAR. TUE
请注意,日期编号和名称/缩写完全不同,因此尝试将它们与固定值进行比较-无论是in (1, 3, 4, 7)
还是使用字符串文字的日期名称-都不可靠。
将日期语言强制为英语可确保比较安全。 (当然,也可以是任何其他语言-字符串文字值只需与您为to_char()
的第三个参数选择的语言匹配即可。)