UPDATE Query without WHERE Clause - sql

Can the UPDATE query be used without a WHERE clause? And if so in what conditions?

if you don't use the WHERE clause all the records on the table will be affected

So, I think when you want to update the whole field for some kind of reasons like updating the status of users enrollment to free for all users.
UPDATE users SET status = "free";

The UPDATE statement in SQL is used to update records in the table. We can modify one or multiple records (rows) in a table using UPDATE statement. If you do not use WHERE clause in UPDATE statement, all the records in the table will be updated.

Related

How to identify which records were not updated after executing an update query in oracle?

I am trying to update 100 records in oracle through the following update statement, but when I execute it, it says 90 records updated, whereas I have 100 records in the where clause.
Now how to identify which 10 records (codes) are not updated out of 100 in oracle?
In the following statement, I want to know the code names which were not updated? is there any simple trick to know?
update table1 a set a.column1='Yes'
where a.column2 in ('code1','code2','code3','code4',........,'code100');
You can't really directly from the update; but you could use the same list of values (assuming it's a hard-coded list, not coming from a table) as a collection, and then look for values in the collection that are not in the table:
select *
from table(sys.odcivarchar2list('code1','code2','code3','code4',........,'code100')) t
where not exists (
select null
from table1 a
where a.column2 = t.column_value
);
db<>fiddle with a smaller set to demonstrate the idea.
You could modify your update to only update rows which are not already 'Yes'; and if you did that then you could either look for collection values that don't exist at all, or those which exist but don't need to be updated - in that case, before you actually run the update, of course. db<>fiddle.
odcivarchar2list is a built-in collection type, but you could use your own.
If you already have the values in a collection or table you can use that directly, both for this query and for the update.

deleting a row in legacy SQL

SELECT * FROM [dataset.HistoricalDashboard]
OMIT RECORD IF IncidentNumber = 'INC0887'
I want to delete a row with that Incident number. Does this query deletes the row? Legacy SQL does not have delete function. so trying to figure it out
OMIT RECORD does not delete the row from the table. Instead, it populates whole results except for the records which are satisfying the given condition. You can use the query(given in question) and store results in another table.
For more information, please refer to the google doc:
https://cloud.google.com/bigquery/docs/reference/legacy-sql#omit

BigQuery DML on partitioned table with nested required fields in nullable records

We've got a data lake in BigQuery, which is a daily partitioned table (with several TBs of data / day). The table should have a unique identifier based on a composite key consisting of multiple fields.
Every now and then, we have to do modifications to historical data, which we approached using DML. The goal of the modification, is to have an easy way of modifying the data, that is cost efficient and atomic.
The challenge arises, due to the schema that our table has. I'll illustrate using an example. Note that this table is not partitioned, though that should not matter for the illustration, only for the costs. The final solution should leverage the fact that the table is partitioned.
Assume we have the a table, with the following schema:
And the following data:
Let's say that fruit.color should be changed with a DML statement. Note the fact, that fruit is a nullable record, and fruit.type is required.
The DML can be executed using either an UPDATE or a MERGE query.
Using UPDATE
#standardSQL
UPDATE `test_dataset.fruity_table`
SET
fruit.color = 'unknown'
WHERE id in ('1', '2')
Unfortunately, the UPDATE statement, does not have the option to conditionally update a field. This is required, since the query above fails with Required field fruit.type cannot be null; error in writing field fruit. It is possible to use a subquery for this, but in our schema, required fields in nullable records are very common. Hence, we'd have to do multiple update statements, which would break the atomic requirement.
Using MERGE
#standardSQL
MERGE `test_dataset.fruity_table` t
USING `test_dataset.fruity_table` s
ON t.id = s.id
WHEN MATCHED AND t.fruit IS NOT NULL
THEN UPDATE SET fruit.color = 'unknown'
WHEN MATCHED AND t.fruit IS NULL
THEN -- do nothing, maybe update some other fields that need to be changed
For the MERGE statement, we unfortunately have to deal with the fact that there are duplicates in the BigQuery table. That is something a MERGE statement cannot handle. It fails with UPDATE/MERGE must match at most one source row for each target row.
As you can see, we have the need for the granularity of the MERGE query, with the flexibility of the UPDATE query. If MERGE would just update all the records that are matched, this would work.
Any ideas on how to solve this? Maybe a totally different approach?
Use UPDATE, but rather than SET'ing fruit.color, SET fruit - similar to how you do it in MERGE. You just need to express options you have in MERGE as a single expression setting fruit, either to NULL or something with non-NULL required field:
update test_dataset.fruity_table
SET fruit = IF(fruit IS NOT NULL, STRUCT('unknown', fruit.type), NULL)
where id = '3'

UPDATE query vs. Table FILTER yields different rows count (MS ACCESS)

I have a problem with count of updated rows through UPDATE query.
When I run UPDATE query:
UPDATE t002Vyk
INNER JOIN tZemnyPlyn ON t002Vyk.BMsID = tZemnyPlyn.[ID budovy]
SET t002Vyk.isZemnyPlynHistoricky = Yes;
it says, that 2916 rows will be updated.
I run the UPDATE query. (I tried different queries as well, with same results).
When I then filter the table t002Vyk ONLY on field isZemnyPlynHistoricky = YES, I only get 2701 rows.
Just to provide all info, here is the design view of field isZemnyPlynHistoricky:
(I also tried using number data type, with 0/1 values, no luck)
My questions
Why do I get different results in filtered table, than in UPDATE query? How to solve the problem?
Solution:
In the tZemnyPlyn table, there were duplicate ID's [ID budovy], which were counted more than once time in UPDATE query results.
After creating a unique list of [ID budovy] the UPDATE query showed 2701 rows as well as the table did.
Thanks to Tim Biegeleisen's comment, for getting me on track.

Sql Query Select From an update, records, that were not updated

I'm using Sql server 2012
Just curious if there is a way to do such a thing...
My update query looks like
UPDATE a
SET a.TowerNumber=b.SiteNumber
FROM tower a
INNER JOIN sites b
ON a.sitenumber = b.sitenumber
what i would like to do after the ON line is something like
EDIT **select records that did not match the ON statement**
thanks in advance
This query returns all rows that wouldn't be touched by the update because the ON (join) criteria wasn't met.
SELECT *
FROM tower
WHERE sitenumber NOT IN
(SELECT sitenumber
FROM sites)
I think you can try to use an INSTEAD OF trigger and then make use of the DELETED table. It should contain the rows that were updated during the update operation so what you are looking after are the rest of the rows in the initial table (select rows from your table where not in DELETED table)