I have written a sql query:
WITH
type = "id" AND match(value, '*** ID WAS ACTIVATED ***') as id_activ
SELECT
multiIf(
id_activ, 'id_activ',
NULL) as case,
date,
value
FROM my.data.base
However it doesn't work. A problem occurs because of this part: match(value, '*** ID WAS ACTIVATED ***'). When i try this, it works perfectly:
WITH
type = "id" AND value LIKE '%*** ID WAS ACTIVATED ***%' as id_activ
SELECT
multiIf(
id_activ, 'id_activ',
NULL) as case,
date,
value
FROM my.data.base
Here is a full message with that pattern that I use in match:
*** ID WAS ACTIVATED *** : Values expired
How could i fix it and write query with match right?
Same thing with query which has square brackets in it:
WITH
type = "id" AND match(value, 'Values processed [NUMBERS]') as id_number
SELECT
multiIf(
id_number, 'id_number',
NULL) as case,
date,
value
FROM my.data.base
match() is matching regular expression
"Equivalent" to
LIKE '%*** ID WAS ACTIVATED ***%'
is
match(value, '.*\*\*\* ID WAS ACTIVATED \*\*\*.*')
https://regex101.com/r/XkRmEY/1
Related
I have a table that has a column video_ids, it is of a bigint[] type. I would like to find all the rows that have any of the elements from the array passed in a select statement. So, if I have a row that has a video_ids field that looks like this:
{9529387, 9548200, 9579636}
I would like to fetch it if I pass an array that has any of this video_ids. I thought I would do that with any, but I am not sure how to do this in SQL, I have tried with this:
select id, finished, failed, video_ids, invoiced_video_ids, failed_video_ids
from video_order_execution
where order_ids = any(
'{9548200, 11934626, 9579636, 11936321, 11509698, 11552728, 11592106, 11643565, 11707543, 11810386, 11846268}'
::bigint[]);
I get an error if I do that:
ERROR: operator does not exist: bigint[] = bigint Hint: No operator
matches the given name and argument types. You might need to add
explicit type casts.
How can I make such a statement that would do the job for what I need?
Use the operator && which returns true if the 2 operands have any common items:
select id, finished, failed, video_ids, invoiced_video_ids, failed_video_ids
from video_order_execution
where order_ids &&
'{9548200, 11934626, 9579636, 11936321, 11509698, 11552728, 11592106, 11643565, 11707543, 11810386, 11846268}'::bigint[];
A proprietary third-party application stores JSON strings in it's database like this one:
{"state":"complete","timestamp":1614776473000}
I need the timestamp and found out that
DB2 offers JSON functions. Since it's stored as string in the PROF_VALUE column, I guess that converting with SYSTOOLS.JSON2BSON is required, before I can use JSON_VAL to fetch the timestamp:
SELECT SYSTOOLS.JSON_VAL(SYSTOOLS.JSON2BSON(PROF_VALUE), "timestamp", "f")
FROM EMPINST.PROFILE_EXTENSIONS ext
WHERE PROF_PROPERTY_ID = 'touchpointState'
This causes an error that timestamp is invalid in the used context ( SQLCODE=-206, SQLSTATE=42703, DRIVER=4.26.14). The same error is thown when I remove the JSON2BSON call like this
SELECT SYSTOOLS.JSON_VAL(PROF_VALUE, "timestamp", "f")
Also not working with the same error (different data-types):
SELECT SYSTOOLS.JSON_VAL(SYSTOOLS.JSON2BSON(PROF_VALUE), "state", "s:1000")
SELECT SYSTOOLS.JSON_VAL(PROF_VALUE) "state", "s:1000")
I don't understand this error. My syntax is like the documented JSON_VAL ( json-value , search-string , result-type) and it is the same like in the examples, where they show how to fetch the name field of an object.
I also played around a bit with JSON_TABLE to use raw input data for testing (instead of the database data), but it seems not suiteable for that.
SELECT *
FROM TABLE(SYSTOOLS.JSON_TABLE( SYSTOOLS.JSON2BSON('{"state":"complete","timestamp":1614776473000}'), 'state','s:32')) DATA
This gave me a table with one row: Type = 2 and Value = complete.
I had two problems in my query: First it seems that double quotes " are for object references. I wasn't aware that there is any difference, because in most databases I used yet, both single ' and double quotes " are equal.
The second problem is, that JSON_VAL needs to be called without SYSTOOLS, but the reference is still needed on SYSTOOLS.JSON2BSON(PROF_VALUE).
With those changes, the following query worked:
SELECT JSON_VAL(SYSTOOLS.JSON2BSON(PROF_VALUE), 'timestamp', 'f')
FROM EMPINST.PROFILE_EXTENSIONS ext
WHERE PROF_PROPERTY_ID = 'touchpointState'
I have a procedure with the parameter IT_ATINN:
IMPORTING
REFERENCE(IT_ATINN) TYPE STRING_TABLE
IT_ATINN contains a list of characteristics.
I have the following code:
LOOP AT values_tab INTO DATA(value).
SELECT ( #value-INSTANCE ) AS CUOBJ
FROM IBSYMBOL
WHERE SYMBOL_ID = #value-SYMBOL_ID
AND ATINN ??? "<======== HERE ???
APPENDING TABLE #DATA(ibsymbol_tab).
ENDLOOP.
How can I check if ATINN (in the WHERE clause) is equal to any entry in IT_ATINN?
To achieve what you want (and I assume you want dynamic SELECT fields) you cannot use inline declarations here, both in LOOP and in SELECT:
The structure of the results set must be statically identifiable. The SELECT list and the FROM clause must be specified statically and host variables in the SELECT list must not be generic.
So either you use inline or use dynamics, not both.
Here is the snippet that illustrates Sandra good suggestion:
TYPES: BEGIN OF ty_value_tab,
instance TYPE char18,
symbol_id TYPE id,
END OF ty_value_tab.
DATA: it_atinn TYPE string_table.
DATA: rt_atinn TYPE RANGE OF atinn,
value TYPE ty_value_tab,
values_tab TYPE RANGE OF ty_value_tab,
ibsymbol_tab TYPE TABLE OF ibsymbol.
rt_atinn = VALUE #( FOR value_atinn IN it_atinn ( sign = 'I' option = 'EQ' low = value_atinn ) ).
APPEND VALUE ty_value_tab( instance = 'ATWRT' ) TO values_tab.
LOOP AT values_tab INTO value.
SELECT (value-instance)
FROM ibsymbol
WHERE symbol_id = #value-symbol_id
AND atinn IN #rt_atinn
APPENDING CORRESPONDING FIELDS OF TABLE #ibsymbol_tab.
ENDLOOP.
Overall, it makes no sense select ibsymbol in loop, 'cause it has only 8 fields, so you can easily collect all necessary fields from values_tab and pass them as dynamic fieldstring.
If you wanna use alias CUOBJ for your dynamic field you should add it like this:
LOOP AT values_tab INTO value.
DATA(aliased_value) = value-instance && ` AS cuobj `.
SELECT (aliased_value)
...
Remember, that your alias should exists among ibsymbol fields, otherwise in case of static ibsymbol_tab declaration this statement will throw a short dump.
I have a Field Field<T>. I want to create a named value for that field, to be able to use it in a query. The name of the value should be the name of the field.
select value as field from ...
Is the the correct way to do it?
public <T> Field<T> namedValue(Field<T> field, T value) {
return DSL.val(value, field).as(field);
}
Although it works, I was wondering if there is a shorter way to do this. I might be pedantic here :).
update
I am creating the following construction:
UPADTE table SET x = alias.x, y = alias.y
FROM (SELECT constant value for x, table2.y FROM table2 WHERE ...) AS alias.
Let's simplify this to (for the sake of this example, to focus on the constant selection):
SELECT
FROM (SELECT constant value for x) AS alias.
First, I started with:
Select s1 = context.select(DSL.val("TEST"));
Select s2 = context.select(s1.fields()).from(s1);
This resulted in an incorrect query:
select "alias_66794930"."TEST" from (select 'TEST') as "alias_66794930"
(I am not really sure if this is correct behavior from jOOQ.)
So, I added an alias:
Select s1 = context.select(DSL.val("TEST").as(X));
Select s2 = context.select(s1.fields()).from(s1);
This resulted in:
select "alias_76324565"."x" from (select 'TEST' as "x") as "alias_76324565"
This works fine. Then, I ran into problems when the constant vale was null:
Select s1 = context.select(DSL.val(null).as(X));
Select s2 = context.select(s1.fields()).from(s1);
This resulted in:
select "alias_85795854"."x" from (select cast(? as varchar) as "x") as "alias_85795854"
1400 [localhost-startStop-1] TRACE org.jooq.impl.DefaultBinding - Binding variable 1 : null (class java.lang.Object)
This makes sense, the field type is not known. So I added the field (with its type) as following:
Select s1 = context.select(DSL.val(null, X).as(X));
Select s2 = context.select(s1.fields()).from(s1);
Binding is now correct:
1678 [localhost-startStop-1] TRACE org.jooq.impl.DefaultBinding - Binding variable 1 : null (class java.lang.String)
All done!
I don't think you can get much shorter than what you already have. I mean, your SQL reads:
value as field
And your Java/jOOQ code reads:
DSL.val(value, field).as(field)
You could of course static import DSL.val or DSL.*:
import static org.jooq.impl.DSL.*;
And then shorten things to:
val(value, field).as(field)
And if you're very sure about value's type, you don't need to coerce it to that of field
val(value).as(field)
Now, you definitely can't go any shorter, and there's no more need for your namedValue() function...
I created a view as below from my existing tables:
Create View LostCase
As
Select L.LNO,Count(*) "NoOfCasesLost"
From L, LWCS,C
Where L.LNO=LWCS.LNO
And LWCS.CASENO=C.CNO
And C.OUTCOME='LOSE'
Group By L.LNO ;
However, whenever I run query as per below , I always get
invalid column name for LostCase.NoOfCasesLost
Select L.LNO,L.LNAME,LostCase.NoOfCasesLost
From L, LostCase
Where L.LNO = LostCase.LNO;
I don't get it why it happens.
If you have specified the column name in the quotation marks, you have to refer it afterwards as:
Select L.LNO,L.LNAME,LostCase."NoOfCasesLost"
From L,LostCase
Where L.LNO=LostCase.LNO;