Updating a column that is part of your selection criteria should be no problem I think/thought.
I still have a query that gives odd results:
update table1 as t1
inner join table2 as t2 on t1.id = t2.old_id
set t1.id = t2.id
I use table2 to map id to old_id.
table1.id and table2.id are both primary keys. table2.old_id is also unique.
It follows that table1.id will still be unique after this update.
WRONG! MS Access will make rumble of this, with some duplicate table.id values.
I think however this is the correct way of updating a column that is used in a join? How could we achieve the desired result in MS Access?
Note: table2 is a mysql view accessed via ODBC.
table1.id and table2.id are both
primary keys. table2.old_id is also
unique.
It follows that table1.id will still
be unique after this update.
That ain't necessarily so.
It seems old_ID and (new) ID are of the same data type. A row in table1 that has no matching row in table2 based on the predicate (table1.ID = table2.old_ID) would not be updated. A different row could match a (new) ID value, hence get updated, with the same value as that row that didn't get updated.
Here's an example using Standard SQL (works in SQL Server 2008, not it Access/Jet) which I hope you can follow. Note I've re-written your INNER JOIN using an EXISTS construct to fit the logic of what I am trying to convey:
WITH Table1 (ID) AS
(
SELECT ID
FROM (
VALUES (1),
(2),
(3),
(4)
) AS Table2 (ID)
),
Table2 (old_ID, ID) AS
(
SELECT old_ID, ID
FROM (
VALUES (1, 55),
(2, 99),
(3, 4)
) AS Table2 (old_ID, ID)
)
-- ID of rows that will not be updated:
SELECT T1.ID
FROM Table1 AS T1
WHERE NOT EXISTS (
SELECT *
FROM Table2 AS T2
WHERE T1.ID = T2.old_ID
)
UNION ALL
-- updated IDs
SELECT T2.ID
FROM Table2 AS T2
WHERE EXISTS (
SELECT *
FROM Table1 AS T1
WHERE T1.ID = T2.old_ID
);
The resultset:
ID
---
4 --<-- duplicate
55
99
4 --<-- duplicate
In other words, even though all the following are unique:
(table1.ID)
(table2.ID)
(table2.old_ID)
...the following may contain duplicates:
table1.ID
UNION ALL
table2.ID
Using JOINs in UPDATE statements isn't consistently supported, which is why they aren't my habit to use.
UPDATE TABLE1 AS t1
SET t1.id = (SELECT t2.id
FROM TABLE2 t2
WHERE t2.old_id = t1.id)
Related
I am trying to update a table in SQL Server. Table is called table1 with columns(code, desc). I want to update table1 with unique records from table2(code, desc). table2 is a copy of table1 but it contains records that are not present in table1.
EDIT 1: table1 consists of all the records in our ERP when exported in July, table2 consists of all the records in our ERP when exported in November (including records added after July which are the records that I want to add to table1)
To update table1.desc with values from matching rows in table2 simply do:
update t1 set
t1.desc = t2.desc
from table1 t1
join table2 t2 on t2.code = t1.code;
If however you want to insert rows into table1 that only exist in table2 (it's not exactly clear if that's the case) you can use not exists
insert into table1 (code, desc)
select code, desc
from table2 t2
where not exists (select * from table1 t1 where t1.code = t2.code);
Sounds like you want an INSERT
Something like this should work:
INSERT INTO table1 (code, desc)
SELECT t2.code, t2.desc
FROM table2 t2
LEFT JOIN table1 t1 on t2.code = t1.code and t1.desc = t2.desc
WHERE t1.code is null --Ignore records that already exist in table1
... Adjust join clause accordingly.
I have two tables: table1 and table2:
table1 has columns id and integer
table2 has columns id and boolean
table2 can have multiple rows with the same id
I want to update the integer column of table1 by looking at all rows with the same id in table2 and seeing if any of the boolean values are true. If so I want table1.integer to be 1, else I want it to be 0.
I have tried something like this:
UPDATE table1,
(
SELECT table2.id, Sum(table2.boolean) > 0
) AS 'condition'
from table2
WHERE 1
GROUP BY table2.id) table3
SET table1.integer =IF(table3.condition, 1, 0) where table1.id = table3.id
And it seems to work, but I wanted to ask if there is a nicer/cleaner/more succinct way of updating the rows of table1 according to multiple rows of table2.
I would recommend EXISTS:
UPDATE table1 t1
SET t1.integer = (EXISTS (SELECT 1
FROM table2 t2
WHERE t2.id = t.id AND
t2.boolean
)
);
This can take advantage of an index on table2(id, boolean). With such an index, it should be faster than an approach that uses JOIN and AGGREGATION.
The syntax of your query is MySql like, so you can do a join like this:
UPDATE table1 t1 INNER JOIN (
SELECT id, MAX(boolean) maxboolean
FROM table2
GROUP BY id
) t2 ON t2.id = t1.id
SET t1.integer = t2.maxboolean
If there are ids in table1 without a corresponding id in table2 and you want the integer column for them to be updated to 0 then use a LEFT join:
UPDATE table1 t1 LEFT JOIN (
SELECT id, MAX(boolean) maxboolean
FROM table2
GROUP BY id
) t2 ON t2.id = t1.id
SET t1.integer = COALESCE(t2.maxboolean, 0)
I have two tables table1 and table2. Those tables have unique name and id columns.
I also have a relation / join table, table1_table2 which has the straight-forward colums table1_id and table2_id.
What I want to do is to insert a new relation into table1_table2 knowing the names of the elements in table1 and table2 I want to create a relation for. But I need to get their ids to insert them into table_table2.
What I want is something like that:
insert into table1_table2 values ((select id from table1 where name = 'some_name'), (select id from table2 where name = 'another_name'))
I also tried using
insert into table1_table2 values ((select id from (select id from table1 where name = 'some_name') where rownum=1), (select id from (select id from table2 where name = 'another_name') where rownum=1))
which also didn't work.
I understand I can first extract the ids if necessary but I'd prefer it to be in one statement.
Edit: I've also tried
insert into table1_table2 values (select t1.id, t2.id from table1 t1, table2 t2 where t1.name = 'some_name' and t2.name = 'another_name')
which also didn't work
Example data:
table1
id name
1 foo
2 bar
table2
id name
1 some
2 data
table1_table2
table1.id table2.id
1 1
and now I want to insert
table1.id table2.id
2 2
into table1_table2, but I do only know that the entry in table1 has the name bar and the entry in table2 has the name data.
This should work:
INSERT INTO table1_table2 (table1_id, table2_id)
VALUES ( (SELECT id FROM table1 WHERE name = 'some_name'),
(SELECT id FROM table2 WHERE name = 'another_name')
);
However, I would write it as:
INSERT INTO table1_table2 (table1_id, table2_id)
SELECT t1.id, t2.id
FROM table1 t1 JOIN
table2 t2
ON t1.name = 'some_name' AND
t2.name = 'another_name';
Note in this case if there is no match in either table, then no row is inserted at all. Using VALUES, a NULL values would be inserted.
I have two differents tables. The have some columns in common, for this example lets say 'name' and 'id'.
By making
( SELECT name,id FROM table1
EXCEPT
SELECT name,id FROM table2)
UNION ALL
( SELECT name,id FROM table2
EXCEPT
SELECT name,id FROM table1)
I get a list of the elements that are on one tablet but not in the other one.
Up to this point everything is OK.
But now, I want to make a select all from table1 where the name and the id matches the result of the query above.
After lots of comments I think this is what you're after...
SELECT T1.*
FROM table1 t1
LEFT JOIN table2 t2
on T1.ID = T2.ID
and T1.Name = T2.Name
AND E2.event_Time_UTC between convert(datetime,'2016-02-09 00:00:20',101) and convert(datetime '2016-02-09 23:59:52',101)
WHERE T2.Name is null
AND E1.Event_Time_UTC between convert(datetime,'2016-02-09 00:00:20',101) and convert(datetime,'2016-02-09 23:59:52',101)
You may allow implicit casting to work but above is the explicit approach.
If not then you would need to cast the string dates to a date time, assuming Event_Time_UTC is a date/time datatype.
A left join lets us return all records from the 1st table and only those that match from the 2nd.
The t1.* returns only the columns from table1. The join criteria (on) allows us to identify those records which match so they can then be eliminated in the where clause by 'where t2.name is null' they will always be null when no record match in t2.
Thus you get a result set that is: all records from t1 without a matching record on name and id in table2.
Old version
The below content is no longer relevant, based on comments.
I redacted previous answer a lot because you're using SQL Server not MySQL and I know you want multiple records not table1 and table2 joined.
In the below I create two tables: table1 and table2. I then populate table1 and table2 with some sample data
I then show how to get only those records which exist in one table but not the other; returning a separate ROW for each. I then go into detail as to why I choose this approach vs others. I'll finally review what you've tried and try to explain why I don't think it will work.
create table table1 (
ID int,
name varchar(20),
col1 varchar(20),
col2 varchar(20),
col3 varchar(20));
Create table table2 (
id int,
name varchar(20));
Insert into table1 values (1,'John','col1','col2','col3');
Insert into table1 values (2,'Paul','col1','col2','col3');
Insert into table1 values (3,'George','col1','col2','col3');
Insert into table2 values (1,'John');
Insert into table2 values (4,'Ringo');
Option 1
SELECT T1.name, T1.ID, T1.Col1, T1.Col2, T1.Col3
FROM Table1 T1
LEFT JOIN Table2 T2
on T1.Name = T2.Name
and T1.ID = T2.ID
WHERE T2.ID is null
UNION ALL
SELECT T2.name, T2.ID, NULL, NULL, NULL
FROM Table1 T1
RIGHT JOIN Table2 T2
on T1.Name = T2.Name
and T1.ID = T2.ID
WHERE T1.ID is null ;
which results in...
Notice John isn't there as it's in both tables. We have the other 2 records from table1, and the ID, name from table2 you're after.
Normally I would do this as a full outer join but since I think you want to reuse the name and id fields to relate to BOTH tables in the same column we had to use this approach and spell out all the column names in table 1 and put NULL for each column in table1 when displaying records from table2 in order to make the output of the second query union to the first. We simply can't use *
Option 2: Using a full outer join... with all columns from T1
SELECT T1.*
FROM Table1 T1
FULL OUTER JOIN Table2 T2
on T1.ID = T2.ID
and T1.Name = T2.Name
WHERE (T1.ID is null or T2.ID is null)
you get this... which doesn't show Ringo...
But then I would ask why you need anything from Table 2 at all so I think you're wanting to still show the ID, Name from table2 when it doesn't exist in table1.
Which is why What I think you're after is the results from the 1st query using the union all.
Option 3 I suppose we could avoid the second query in option 1 by doing...
SELECT coalesce(T1.Name, T2.name) as name, coalesce(T1.Id,T2.ID) as ID, T1.col1, T1.Col2, T1.Col3
FROM Table1 T1
FULL OUTER JOIN Table2 T2
on T1.ID = T2.ID
and T1.Name = T2.Name
WHERE (T1.ID is null or T2.ID is null)
which gives us what I believe to be the desired results as well.
This works because we know we only want the name,id from table2 and all the column values in table1 will be blank.
Notice however in all cases we simply can't use Tablename.* to select all records from table1.
This is what you tried:
( SELECT name,id FROM table1
EXCEPT
SELECT name,id FROM table2)
UNION ALL
( SELECT name,id FROM table2
EXCEPT
SELECT name,id FROM table1)
Assuming you want to reuse the ID, Name fields; you can't select *. Why? because the records in Table2 not in table1 aren't in table1. In my example if you want Ringo to show up you have to reference table2! Additionally, * gives you no ability to "Coalesce" the ID and name fields together as I did in option 3 above.
If you ONLY want the columns from table1, that means you will NEVER see data from table2. If you don't need the data from table2, (such as ringo in my example) then why do we need to do the union at all?) I'm assuming you want ringo, thus you HAVE to somewhere reference name, id from table2.
You could also do this with NOT EXISTS:
SELECT *
FROM table1
WHERE
NOT EXISTS
(SELECT 1
FROM table2
WHERE table1.id = table2.id
AND table1.name = table2.name)
;with cte as
(
( SELECT name,id FROM table1
EXCEPT
SELECT name,id FROM table2)
UNION ALL
( SELECT name,id FROM table2
EXCEPT
SELECT name,id FROM table1)
)
Select *
from table1 as tbl1
where
tbl1.id = cte.id
and tbl1.name = cte.name
I have a SQL Insert statement that needs to insert records into another table only if the the record doesn't exist in table2 or the zip code has changes in table1. I have tried the following but it throws an error and it is the logic I am looking for:
INSERT INTO table2
SELECT id, zip
FROM table1 t1
JOIN table2 t2
ON t1.id = t2.id and t1.zip <> t2.zip
I also need it to insert the records if the id doesn't exist at all in table2. I have googled the crap out of this and can't seem to find the solution anywhere.
What about this?
INSERT INTO table2
SELECT t2.id, t2.zip
FROM table1 t1
LEFT OUTER JOIN table2 t2
ON t1.id = t2.id
WHERE (t1.id IS NULL OR t2.zip <> t1.zip)
Also, be sure to clarify which table's id and zip columns you are asking for.
You should always include column lists when doing inserts. Second, your query doesn't quite capture your logic. You need a left outer join to find the records that don't exist in the second table. Perhaps this might do what you want:
INSERT INTO table2(id, zip)
SELECT id, zip
FROM table1 t1 LEFT JOIN
table2 t2
ON t1.id = t2.id
WHERE (t1.zip <> t2.zip) or (t2.zip is null)
You just need a WHERE NOT EXISTS clause
INSERT INTO table2
SELECT id, zip
FROM table1
WHERE NOT EXISTS (SELECT 1 FROM table2 WHERE table2.id = table1.id AND table2.zip = table1.zip)