Cannot propagate row_count from __TABLES__ metadata table to custom table - google-bigquery

Today I am no longer able to create a table in bigquery from metadata table TABLES. Yesterday everything worked fine. Issue can be reproduced using:
create or replace table `steady-vine-203410.analysis_us.test`
as
select row_count from `bigquery-public-data.usa_names.__TABLES__`
The problem is that during table creation the column row_count is somehow missing. If I try only select without creating the table, everything works.
I have tried doing nested (derived) tables and using WITH clausule, but none of it seems to work today.
Did I miss some bigquery update? Can someone explain to me whats happening or what is the alternative to get row_count of all the tables in a dataset?
Thank you

Related

Auto refresh hive clone table

I am working with the hive, I have created a table by using fields from a master tables.
Sample code
create table final_table select feild1.feild2 from mater_table;
I need to auto update my table when master table is updated.
any suggestions pelase, thanks in advance.
I don't get any solution for this. Anyway, i implemented same with the help of view.
create view final_view select feild1.feild2 from mater_table;
Its ok for my requirement.

Retrieve Script used in "Create Table As" Statement

We have a table in our Oracle Database that was created from an actual script.
Ex:
Create Table AS (Select * from table).
I was hoping to recover the original script the table was created from as the data is quite old in the table, but needs this created table needs to be refreshed. This table is created with data from another live table in our database, so if there is a way to refresh this without the original query - I'm open ears. Any solutions are welcomed!
Thanks!
I suppose you could also do a column by column comparison of this table against all others to see which one (if any) matches it. Of course, this would only be a guess.
It would require that object to actually be a materialized view instead of a table. Otherwise you are probably left off with exploring logs. Beyond that I doubt there is any way to recover the original select statement used to create that table.

SQL old table still appearing

This is a really strange question. I have a dynamic SQL stored procedure that inserts data into a static table in the queried database. This table is referenced quite a lot in the query. So when I needed to change this table and add two new columns I deleted it and used the import wizard (Excel spread sheet) to create a new one and gave it that same name, so I didn't have to amend the SP. The SP works fine, however I also have this query outside of dynamic SQL and when I run it, it now fails.
At first I couldn't work out why but when I saw that it was failing on the INSERT INTO the newly created (but with the same name) table because there were too many columns to match the table. I ran a simple SELECT * FROM and it brought back the old table with the 3 columns it used to have and not the new table with 5 columns?
How can this table still exist if its been deleted? Its like a ghost table still remains?
Thanks
First check your temp table exists or not in your database.IF exists means drop your temp table and then create new one.
IF EXISTS(SELECT 1 FROM tempdb.dbo.sysobjects WHERE xtype in ('U') AND id =
object_id(N'tempdb..#your_tableName') )
DROP TABLE #your_tableName;
This is known behaviour with views and can be fixed by dropping and recreating the view. Not sure why it is acting the same way for a table. Possibilities I can think of -
It is a view
The new table was not created in the same database
Try dropping and recreating the sproc, for what it is worth

Update a Temporary Table before showing?

I'd like to find out how to update a temporary table before I show the query. This is to avoid making permanent changes to the database.
So far I got the following:
WITH
new_salary AS
(SELECT ID,NAME,DEPT_NAME,SALARY FROM INSTRUCTOR WHERE DEPT_NAME='Comp. Sci.')
SELECT
*
FROM
new_salary
WHERE
DEPT_NAME='Comp. Sci.';
Now here is where it ends. I want to update this temporary table and show the updated version of that table as to avoid changing the actual database. All my attempts at using the UPDATE clause have failed so I am kind of dumbfounded :/
This part I am currently trying to do is not part of homework. It's just me who doesn't want to have to re-do the database over and over.
How would I go about doing this?
I guess you have two options:
You make a procedure, which first checks whether it needs to update the table. After calling that you execute the query.
You create a pipelined function, which does the checking and returning of the data. You could integrate this into the select like this (pipelined function name called pipelined_function_name):
select *
from table(pipelined_function_name)
;

How to deal with a Firebird table that's apparently both there and not there

I've just had something very strange happen to me with a Firebird database.
I was trying to create a table, and the CREATE TABLE failed for some reason. But now it's stuck in a very strange state:
If I try to CREATE TABLE again with the same table name, it gives an error: the table already exists. But if I try to DROP TABLE that table, it gives an error: the table does not exist. Trying to SELECT * FROM that table gives the "table does not exist" error, and the name does not show up in the metadata query:
SELECT RDB$RELATION_NAME
FROM RDB$RELATIONS
WHERE RDB$SYSTEM_FLAG=0
So for some reason, the table really seems to not be there, but I can't create it because something somewhere indicates that it does exist.
Does anyone have any idea how to fix this? I've already tried closing all connections to that database, which has helped with inconsistency issues in the past, but this time it doesn't help.
You didn't give details about what was the error when you tried to create the table, so I cannot comment it. But RDB$RELATIONS is not the only system table affected when you create a table. Maybe you are now in an inconsistent situation where some information about that table exists in some system tables and doesn't exists in others.
Another option is corrupted indexes in the system tables, so the record is not there but the index think it still exists.
Try to do a backup/restore and see if it helps. It it doesnt work, try to search for records related to that "non created" table in the other system tables (RDB$RELATION_FIELDS, etc) and if you find any, try to delete them.
As a last option, you may create a new clean database with correct metadata and pump your data to it using IBDataPump.