I'm having some difficulties in the usage of the CONNECT BY statement with Oracle. I have no problem in writing the query for 1 given record (using START WITH) but I'm looking for a possibility to create an extract.
Data
1
2
3
4
5
6
Table
ID PARENT_ID
1 2
1 5
1 6
2 3
2 4
What I need is a table which would be filled with the following information/records.
Desired Output
ID PARENT_ID
1 2
1 3
1 4
1 5
1 6
2 3
2 4
If I use the following query
SELECT PARENT_ID
FROM TABLE
START WITH ID = 1
CONNECT BY NOCYCLE PRIOR ID = PARENT_ID
I will have the following result
2
3
4
5
6
This is effectively the parent_id's which I am looking for for ID 1.
If I remove the start with I have the following list
2
3
4
5
6
3
4
I've tried to add the ID in the statement but that did not work either. Any idea if it's possible to have such an output?
You forgot the prior keyword. You want to connect the parent_id to the id of the previous record. (Could be the other way around, but you'll notice that fast enough. ;-) )
SELECT ID
FROM TABLE
START WITH ID = 1
CONNECT BY prior ID = PARENT_ID
Related
View of a table
ID
kWh
1
3
1
10
1
8
1
11
2
12
2
4
2
7
2
8
3
3
3
4
3
5
I want to recive
ID
kWh
1
32
2
31
3
12
The table itself is more complex and larger. But the point is this. How can this be done? And I can't know in advance the ID numbers of the first column.
SELECT T.ID,SUM(T.KWH)SUM_KWH
FROM YOUR_TABLE T
GROUP BY T.ID
Do you need this one?
Let's assume your database name is 'testdb' and table name is 'table1'.
SELECT * FROM testdb.table1;
SELECT id, SUM(kwh) AS "kwh2"
FROM stack.table1
WHERE id = 1
keep running the query will all (ids). you will get output.
By following this query you will get desired output.
Hope this helps.
I have a large table (3 million rows and about 12 columns). I have one column that can contain duplicate values - this is my "ID" column. I have a second column "NUM_ID" that I would like to have it start at the value of 1 for every unique "ID". Then - if I run into a duplicate value - "NUM_ID" would then bump up one value (to 2) and so on. For example:
ID NUM_ID
1 1
2 1
2 2
2 3
3 1
3 2
4 1
5 1
5 2
5 3
5 4
Again, "ID" is pre-populated, I cannot change this column and its values. My "NUM_ID" column is currently empty - I'm hoping there is a sql command I can use to populate the column as shown above? I've tried using Python but updating 3M rows is taking a long time. Also, if it matters, I am using PostGresSQL.
Help? Thanks!
If you are Using SQL Server then You Should Use ROW_NUMBER() as below :
SELECT *, ROW_NUMBER() OVER (PARTITION BY ID ORDER BY ID) NUM_ID FROM #TM
Result :
ID NUM_ID
1 1
2 1
2 2
2 3
3 1
3 2
4 1
5 1
5 2
5 3
5 4
I have 1 table, which refers by value to rows in the same table
Example table:
ID PARENT_ID NAME
1 0 john
2 1 jane
3 2 smigy
4 2 gujo
5 1 duby
6 1 ruby
7 5 foo
8 2 bar
9 3 baz
10 3 qux
The root-parent has parent 0 (just so it wouldn't be null), in this case there's
1 root-parent - parent(0)=1.
root-parent has 1lvl children - parent(1)=2;5;6.
1lvl children has 2lvl children - parent(2)=3;4;8. parent(5)=7. parent(6) has nothing.
2lvl children has 3lvl children - parent(3)=9;10. parent(4) has nothing. parent(8) has nothing.
There is no lvl4 children or anything with depth beyond 4.
And I need to create a script (presumably SQL query - need to avoid function/procedure/etc.) that would index rows based on their position under their parent.
Just like if I'd select all root-parent's and get (rownum-1)
The goal table should look like this:
ID PARENT_ID NAME ROW_INDEX
1 0 john 0
2 1 jane 0
3 2 smigy 0
4 2 gujo 1
5 1 duby 1
6 1 ruby 2
7 5 foo 0
8 2 bar 2
9 3 baz 0
10 3 qux 1
I'm planing to add this column and thus the query will be executed only once. I've played by selecting seperate depth rows, but then I don't really know how to count inside/between group by (even if that is possible).
P.S. A better/good column name suggestion would also be very appreciated.
User row_number()
select mt.*, row_number() over(partition by parent_id order by id) - 1 as rn
from MyTable mt
I'm trying to create an SQL statement to match either an id number or a postcode and then assign a new id number
What I want to end up with is ‘newid’ that correctly recognizes that the first four records are the same person (even though the postcode for record 2 is different).
record id postcode newid
--------------------------
1 1 1 1
2 1 2 1
3 1 1 1
4 2 1 1
5 3 3 2
Any suggestions would be appreciated greatly.
Going based on your example:
SELECT RECORD,
(SELECT MIN (ID)
FROM users u2
WHERE users.id IN (u2.id, u2.postcode)
OR users.postcode in (u2.id, u2.postcode)
) AS newid
FROM users
This results with the following data:
RECORD NEWID
------------------
1 1
2 1
3 1
4 1
5 3
Here is the SQLFiddle
I am trying to wrap my head around a problem I hit exporting data from one system to another.
Let's say I have a table like:
id | item_num
1 1
2 1
3 2
4 3
5 3
6 3
I need to add a column to the table and update it to contain an incrementing product_num field based on item. This would be the end result given the above table.
id | item_num | product_num
1 1 1
2 1 2
3 2 1
4 3 1
5 3 2
6 3 3
Any ideas on going about this?
Edit: This is being done in Access 2010 from one system to another (sql server source, custom/unknown ODBC driven destination)
Perhaps you could create a view in your SQL Server database and then select from that in Access to insert into your destination.
Possible solutions in SQL Server:
-- Use row_number() to get product_num in SQL Server 2005+:
select id
, item_num
, row_number() over (partition by item_num order by id) as product_num
from MyTable;
-- Use a correlated subquery to get product_num in many databases:
select t.id
, t.item_num
, (select count(*) from MyTable where item_num = t.item_num and id <= t.id) as product_num
from MyTable t;
Same result:
id item_num product_num
----------- ----------- --------------------
1 1 1
2 1 2
3 2 1
4 3 1
5 3 2
6 3 3