Date query

[Ramaswami]

Hi Folks, 

I want to grab the data from a table whose date is only 2 days ago and ignore if he has more than 1 row in the table. 

For example : the below is the table

unique id	respondent id	class	occupation	created date
11	1	abc	abc	11/8/2020
12	1	efg	o	12/8/2020
13	2	huu	p	11/8/2020
14	2	abc	abc	11/8/2020
15	3	efg	o	11/8/2020
16	3	df	df	9/8/2020

if i run the above table it should return the below : 

13	2	huu	p	11/8/2020
14	2	abc	abc	11/8/2020

I want to grab the people of only 11th Aug and if a guy have created date as 11th aug and some other date i should ignore him. Ideally a guy should have activities only for 11th Aug ( it may be more than 1 row and created by is 11 Aug that's fine) he should not have any more activity on other dates. 

i have created a query like this : 

related sql query : 

SELECT N0.ID FROM tablename N0 WHERE (DateOnly(N0.CREATED_DATE, 'America/Chicago') = AddDays(DateOnly(GetDate(), 'America/Chicago') , -2)) AND NOT ((DateOnly(N0.CREATED_DATE, 'America/Chicago') > AddDays(DateOnly(GetDate(), 'America/Chicago') , -2)) AND (DateOnly(N0.CREATED_DATE, 'America/Chicago') < AddDays(DateOnly(GetDate(), 'America/Chicago') , -2))) LIMIT 10000

it is returning me all the rows "NOT" is not working. 

[Milan_Vucetic]

Hi @Ramaswami

If I have correct understanding you need unique record which occured before two days: basicaly you want to fetch records which happended exactly two day ago and there is no duplicates on it in the whole table no matter on the date.

You may use  one Query node and make the following aggregation: SELECT respondent_id, Count(unique id) as CT, Max(trunc(created_date)) as MX from your_table GROUP BY respondent_id  and then use another Query or Split node to query previous one with condition CT = 1 and MX = two days ago

Regards,

Milan

[Ramaswami]

can we do a count(uniqueid) inside a query component? can you please show me an example

[Milan_Vucetic]

Hi @Ramaswami,

set up like this:

1. Count(@unique_id)
2. Max(DateOnly(@created_date))
3. @respondent_id

Regards,

Milan

[Ramaswami]

Hi @milan

tried like this and it says id should be in group by : 

09/03/2020 6:00:51 AM query PGS-220000 PostgreSQL error: ERROR: column "n0.id" must appear in the GROUP BY clause or be used in an aggregate function LINE 1: ...ATED_DATE, 'America/Chicago')), N0.RESPONDENT_ID, N0.ID FROM...

if i keep id in group by it is giving me all the counts like : So for each respondent it is giving me min date of the unique id which is not what we need : 

here the resp is same but because of group by unique id it is giving me all the records. 

count = 1

resp = 123

min date = 2020-08-30

id = 1290

count =1

resp = 123

min date = 2020-08-23

id = 122345

[Milan_Vucetic]

Hi @Ramaswami

exclude id like this:

Regards,

Milan

[Sukrity_Wadhwa]

Hi @Ramaswami,

Was the proposed solution helpful for you? Please let us know.

Thanks!

[Ramaswami]

Nope