JOIN Multiple Table based on timestamp and another condition

问题

I want to join 3 tables which Table A have date and time in different field that need to be combined first so it can be compared on Table B using field timestamp.

Table A

 userid |  date_in   | check_in
--------+------------+----------
 145    | 2017-01-23 | 08:56:05
 254    | 2017-01-24 | 08:56:54
 202    | 2017-01-25 | 08:53:26
 15     | 2017-01-26 | 08:47:40

Table B

 userid |      checktime      |      sn
--------+---------------------+---------------
 145    | 2017-01-23 08:56:05 | 0345135200184
 254    | 2017-01-24 08:56:54 | 0345135200184
 202    | 2017-01-25 08:53:26 | 0345135200184
 15     | 2017-01-26 08:47:40 | 0345135200184

Table 3

      sn       |    alias
---------------+-------------
 0345135200184 | Alam Sutera

I tried this:

select process.userid, process.date_in, process.check_in, checkinout.checktime, iclock.alias from process 
inner join checkinout on checkinout.checktime=(select cast (date_in as timestamp(0)) + (check_in - time '00:00:00') checktime from process)
inner join iclock on checkinout.sn=iclock.sn
order by userid desc limit 10;

I also tried this:

select checkinout.userid, checkinout.checktime, checkinout.sn from checkinout
inner join lateral
(select distinct userid, date_in, check_in from process where 
date_in=(select checktime::date from checkinout) and 
check_in=(select checktime::time from checkinout)
order by date_in asc limit 1)
process ON true;

Both of them give error: ERROR: more than one row returned by a subquery used as an expression

Expected Result:

 userid |  date_in   | check_in |      checktime     | alias
--------+------------+----------+--------------------+-------
 145    | 2017-01-23 | 08:56:05 | 2017-01-21 09:49:04|Alam Sutera
 254    | 2017-01-24 | 08:56:54 | 2017-01-21 07:57:05|Alam Sutera
 202    | 2017-01-25 | 08:53:26 | 2017-01-21 14:27:00|Alam Sutera
 15     | 2017-01-26 | 08:47:40 | 2017-01-21 06:30:34|Alam Sutera

Can someone help me to solve this? Thank you for the help.

回答1:

By adding date and time field you will get the timestamp which you can compare for join. so you can write your query like below:

select 
t1.userid, t1.date_in, t1.check_in, t2.checktime, t3.alias
from process t1
inner join checkinout t2 on t2. checktime= date_in + check_in and t1.userid=t2.userid
inner join table3 t3 on t2.sn=t3.sn

DEMO

Regarding you error you have mentioned in your question ERROR: more than one row returned by a subquery used as an expression is due to join condition you have used.

回答2:

You can convert your separate Date and Time columns to a Timestamp using the To_Timestamp function as shown below. This should enable you to join TableA and TableB as desired.

select * from tableA a 
join tableB b 
on a.userid = b.userid 
and to_timestamp(a.date_in || ' ' || a.check_in, 'YYYY-MM-DD HH24:MI:SS') = b.checktime

As demonstrated here

来源：https://stackoverflow.com/questions/64314541/join-multiple-table-based-on-timestamp-and-another-condition

标签

sql

postgresql

join

timestamp