We have a custom translator written that extends the SQLServerExecutionFactory, since Teiid 9.1. We have overridden the following translator capabilities to avoid push down of these cases.
supportsUnions=false; supportsExcept=false;
supportsIntersect=false; supportsInLineViews = false;
supportsCaseExpressions = false; supportGroupBy= false;
The following query
SELECT emp.emp_id,emp.emp_name from emp
INNER JOIN dept ON dept.dept_id=emp.dept_id
INNER JOIN company ON comp.company_id=dept.company_id
gets transformed and reaches the translator as
select emp.emp_id,emp.emp_name from emp where emp.dept_id is not null and dept.company_id is not null ORDER BY ##version OFFSET 0 ROWS FETCH NEXT 100 ROWS ONLY;
As you can see there is no join on the transformed query and the dept table seems to be missing, even though the tables column is used in the query.
Is this a defect or are we missing any translator capability that has been introduced since Teiid 13?.
Is this a defect or are we missing any translator capability that has been introduced since Teiid 13?
What version are you seeing this on? If it's 13+, please log an issue at https://issues.redhat.com/projects/TEIID
Without knowing the version and having a query debug plan, it's hard to say exactly what the nature of the issue is from just this.
Related
I'm using QuestDB and SQL for the first time, and I stumbled upon the LATEST_ON syntax used in QuestDB. Can someone explain it's usage and where to use it?
Quoted from the docs:
For scenarios where multiple time series are stored in the same table, it is relatively difficult to identify the latest items of these time series with standard SQL syntax. QuestDB introduces LATEST ON clause for a SELECT statement to remove boilerplate clutter and splice the table with relative ease.
For more information visit the official documentation
LATEST ON is to find the latest record for each unique time series in a table. See this page for some examples: https://questdb.io/docs/reference/sql/latest-on/
It gives you the latest available record for each combination of the PARTITION BY values, according to the ON timestamp
Maybe easier to understand with an example. If you go to https://demo.questdb.io you can execute this query
select * from trades latest on timestamp
partition by symbol, side
It will then show you the latest existing row for each combination of Symbol and Side. If you wanted to do this using standard SQL, you would probably have to use a window function, something like this
select * from
(select *
,ROW_NUMBER() over (partition by Symbol, Side
order by timestamp DESC) AS RowNumber
from trades where timestamp > '2022-10-01') t
where t.RowNumber = 0
Latest on retrieves the latest entry by timestamp for a given key or combination of keys, for scenarios where multiple time series are stored in the same table.
Check this link for some examples: https://questdb.io/docs/reference/sql/latest-on/
I have 2 external tables in BiqQuery, created on top of JSON files on Google Cloud Storage. The first one is a fact table, the second is errors data - and it might or might not be empty.
I can query each table separately just fine, even an empty one - here is an
empty table query result example
I'm also able to left join them if both of them are not empty.
However, if errors table is empty, my query fails with the following error:
The query specified one or more federated data sources but not all of them were scanned. It usually indicates incorrect uri specification or a 'limit' clause over a union of federated data sources that was satisfied without having to read all sources.
This situation isn't covered anywhere in the docs, and it's not related to this versioning issue - Reading BigQuery federated table as source in Dataflow throws an error
I'd rather avoid converting either of this tables to native, since they are used in just one step of the ETL process, and this data is dropped afterwards. One of them being empty doesn't look like an exceptional situation, since plain select works just fine.
Is some workaround possible?
UPD: raised an issue with Google, waiting for response - https://issuetracker.google.com/issues/145230326
It feels like a bug. One workaround is to use scripting to avoid querying the empty table:
DECLARE is_external_table_empty BOOL DEFAULT
(SELECT 0 = (SELECT COUNT(*) FROM your_external_table));
-- do things differently when is_external_table_empty is true
IF is_external_table_empty = true
THEN ...
ELSE ...
END IF
I am trying to save data into my table using an interactive grid with the help of custom plsql. I am running into an "ORA-01403-no data found" error while inserting data and I can't figure out why.
This is my plsql custom process which I run. Appreciate your help.
DECLARE
em_id NUMBER;
BEGIN
CASE :apex$row_status
WHEN 'C'
THEN
SELECT NVL (MAX (emergency_id), 0) + 1
INTO em_id
FROM emp_emergency_contact;
INSERT INTO emp_emergency_contact
(emergency_id, emp_id, emergency_name, emergency_relation
)
VALUES (em_id, :emp_id, :emergency_name, :emergency_relation
);
WHEN 'U'
THEN
UPDATE emp_emergency_contact
SET emergency_name = :emergency_name,
emergency_relation = :emergency_relation
WHERE emergency_id = :emergency_id;
WHEN 'D'
THEN
DELETE emp_emergency_contact
WHERE emergency_id = :emergency_id;
END CASE;
END;
So far I have not come across any documented way on how to use custom PL/SQL logic for processing submitted rows of APEX 5.1 Interactive Grid via AJAX call.
You are getting no data found error because the return is expected to be in certain json format.
The example you have provided is not too complex and can be with done using standard "Interactive Grid - Automatic Row Processing (DML)" process, which is an AJAX approach. If AJAX call is not important then you can create your own PL/SQL process with custom logic. Example of which is demonstrated in "Sample Interactive Grids" package application, check out Advanced > Custom Server Processing page in this application for more information.
I agree with Scott, you should be using a sequence or identity column for ids.
Not entirely sure. A 'select into' can raise a no_data_found exception, but yours shouldn't.
That being said, you shouldn't have max(id)+1 anywhere in your code. This is a bug. Use a sequence or identity column instead.
I have gotten this many times so the first thing I do is go look at any columns in my grid sql that are not part of the "Save", they are from a join for data only.
I just got it again and it was a heading sort column that I had as a column type of "Number". I changed it to display only and the "Save" now works.
Although, I had already set the "Source" of the column to "Query Only" which is also needed.
It is a bummer the Ajax error message doesn't at least give the column name that caused the error.
Hope this helps someone..
BillC
Add a RETURNING INTO clause after the insert. IG expects a primary key to be returned to query the inserted row.
i need to convert this sql query to hibernate criteria,please help guys.
SELECT NAME, COUNT(*) AS app
FROM device
GROUP BY NAME
ORDER BY app DESC
LIMIT 3
Try this code:
select device.name, count(device)
from Device device
group by device.name
order by count(device) desc
This assumes that you have an entity class called Device with a field name along with a getter method getName(). You may have to change the query depending on what your actual code is (which you never showed us).
The LIMIT clause you had is not applicable for HQL. Instead, you should be doing Query.setMaxResults().
I have a Borland builder c++ 6 application calling Oracle 10g database. Operating over a LAN. When the application in question makes a simple db select e.g.
select table_name from element_tablenames where element_id = 10023842
the following is recorded as happening in Oracle (from the performance logs)
select table_name
from element_tablenames
where element_id = 10023842
then immediately (and not from C++ source code but perhaps deeper)
select table_name, element_tablenames.ROWID
from element_tablenames
where element_id = 10023842
The select statement is only called once in the TADODbQuery object, yet two queries are being performed - one to parse and the other adds the ROWID for executon.
Over a WAN and many, many queries this is obviously a problem to the user.
Does anyone know why this might be happening, can someone suggest a solution?
Agree with Robert.
The ROWID uniquely identifies a row in a table so that the returned record can be applied back to the database with any changes (or as a DELETE).
Is there a way to identify a particular column (or set of columns) as a primary key so that it can be used to identify a row without using a ROWID.
I don't know exactly where the RowID is coming from, it could be either the TAdoQuery implementation or the Oracle Driver. But I am sure I found the reason.
From the Oracle docs:
If the database table does not contain a primary key, the ROWID must be selected explicitly when populating DataTable.
So I suspect your Table does not have a primary key, either add one or add the rowid.
Either way this will solve the duplicate query problem.
Since you are concerned about performance. In general
Using TAdoQuery you can set the CursorType to optimize different behaviors for performance. This article covers this from a TAdoQuery perspective. MSDN also has an article that covers it from from a general ADO Perspective. Finally the specifications from the Oracle Driver can be useful.
I would recommend setting the Cursor to either as they are the only supported by Oracle
ctStatic - Bi-directional query produced.
ctOpenForwardOnly - Unidirectional query produced, fastest but can't call Prior
You can also play with CursorLocation to see how it effects your speed.