I have 2 columns in a sql database that are integers (in separate tables), i'll simplify my problem to explain. I have to replicate what has been done in excel in SQL and don't know how to do it. In basic terms I have to loop through all of one table multiplying all by the first record in the 2nd table, then the 2nd row and so on. Building up a table. I don't know how to do this does anyone have any ideas? Example below.
Column A:| Column B:
12 | 36
24 | 89
26
The result output should a table like:
A | B
432 | 1068
864 | 2136
936 | 2314
So 36 has been multiplied by 12, 24 then 26 to create the first column and then 89 multiplied by 12, 24 then 26 to create the 2nd column and so on
The reality columns have 400+ rows to multiply by so the result is a huge table, how can i do this looping through and adding columns to build a calculated table. Hope that makes sense, any help would be greatly appreciated.
Since I dont know how your database looks like here is a generic code that could work for you.
select tableA.colA, tableB.colb, (tableA.colA*tableB.colB) as colC
from tableA
join tableB (on ... if there is something to join on).
This will give you three columns though, the two original columns and the third multiplied column. You could ofcourse remove colA and colB oif you dont need them, they are just there for reference.
this code should work in most sql-languages.
Many thanks for your help, I sorted out the problem with a cross join
Related
Below are the input and output details.Any database Oracle, SQL Server and MySQL should do for the answers.I am not able to derive the logic to rank data which will help me to pivot.
My source is a flat file which contains data like below.I have loaded that file into one of the tables in Oracle.
Source Input:
**Flatfile1**
**Coulmn1**
Kamesh
65
5000
123456789
Nanu
45
3000
321654789
Expected Output:
Name Age Salary Mobilenumber
Kamesh 65 5000 123456789
Nanu 45 3000 321654789
After loading into one of the tables I am applying the logic to number this data which will eventually look like below:
Column1 Datavalue
Kamesh 1
65 1
5000 1
123456789 1
Nanu 2
45 2
3000 2
321654789 2
However, I am not able to derive logic (I tried with Rank) which will give me sequence number like this without having any key field.Hope this explains situation.
Thanks!!
Oracle doesn't store the rows in order, if you do select * from table1 multiple times you could get rows in different orders according to db operations and caching
Therefore if you have a table like that with no other column it's impossible to "pivot" the data.
I strongly suggest to save data in a normalized form, if you can't consider adding a column with a row ID populated automatically (identity column in oracle 12, trigger+ sequence in previous version)
Once you have your rows in order it will be easy to organize your data
What I'm trying to do is add another column to an existing table whose value will depend on an already existing column in the table. For example say I have this table:
Table1
|Letter|
A
C
R
A
I want to create another column (for example, numbers) that is chosen based on the letters. So let's say A corresponds with 10, C with 3 and R with 32 (this was chosen at random). My resulting table should be like this:
|Letter| Number |
A | 10
C | 3
R | 32
A | 10
Can anyone help me write a query that does this..I have over 20 different cases, so the simpler it looks the better.
Thanks in advance!
Options:
Build a table that associates [Letter] with the numeric value. Include this table in query by joining on the common [Letter] fields.
A very long Switch() expression. However, query design grid cell has a limit of 1024 characters.
Better to provide example with your real data and criteria.
Background:
I am working on Sql2000 query where I need to get a cross tab query report. I tried few combinations and got the results using RollUP.
Issue
The real problem is that I have to insert the records in a table which has only one column.
For instance:
My actual records (after RollUP) have 3 columns and are as follow:
ID|May|Jun|Total
CS|21 |54 |75
AB|12 |15 |27
(These results are coming in grid format)
The problem I need to insert in them in a table having only 1 column. Format would be
Data
ID May Jun Total
21 54 75
12 15 27
Please note that column header is also inserted in the second table.
Does this makes sense?
I know this can be done easily with reporing services or crystal reports but I have limiation of not going for them.
Thanks in advance.
It sounds like all you want to do is concatenate the strings of the three (or four?) columns into one?
I'll save you the bad practice lecture that usually accompanies this type of question:
Based on your example, I am assuming you want fixed length fields, and not delimited ones. If you want delimited, use varchar instead of char and put delimiting strings in between. The last column is cast as varchar since you don't need the extra padding.
SELECT CAST([ID] AS char(10))
+ CAST(May AS char(10))
+ CAST(June AS char(10))
+ CAST([Total] AS varchar(10)) AS column1
FROM ...
This would result in
CS 21 54 75
AB 12 15 27
I have three tables.
Table Data contains data for individual parts that come from a
"data.txt" file.
Table Limits contains the limits for the Data table
from a "limits.txt" file.
Table Files is a listing for
each individual .txt file above.
So the "Files" table looks like this. As you can see it is a listing of each file that exists. The LimitsA file will contain the limits for every Data file of type A.
ID File_Name Type Sub-Type
1 DataA_10 A 10
2 DataA_20 A 20
3 DataA_30 A 30
4 LimitsA A NONE
5 DataB_10 B 10
6 DataB_20 B 20
7 LimitsB B NONE
The "Data" table looks like this. The File_ID is the foreign key from the "Files" table. Specifically, this would be data for DataA_10 above:
ID File_ID Dat1 Dat2 Dat3... Dat20
1 1 50 52 53
2 1 12 43 52
3 1 32 42 62
The "Limits" table looks like this. The File_ID is the foreign key from the "Files" table. Specifically, this would be data for LimitsA above:
ID File_ID Sub-Type Lim1 Lim2
1 4 10 40 60
2 4 20 20 30
3 4 30 10 20
So what I want to do is JOIN the correct limits from the "Limit" table to the data from the corresponding "Data" table. Each row of DataA_10 would have the limits of "40" and "60" from the LimitsA table. Unfortunately there is no way to directly link the limits table to the data table. The only way to do this would be to look back to the files table and see that LimitsA and DataA_10 are of type A. Once I link those two together I then need to specifically only grab the Limits for Sub-Type 10.
In the end I would like to have a result that looks like this.
Result:
ID File_ID Dat1 Dat2 Dat3... Dat20 Lim1 Lim2
1 1 50 52 53 40 60
2 1 12 43 52 40 60
3 1 32 42 62 40 60
I hope this is clear enough to understand. It seems to me like an issue of joining more than 2 tables, but I have been unable to find a suitable solution online as of yet. If you have a solution or any advice it would be greatly appreciated.
Your 'Files' table is actually 2 separate (but related) concepts that have been merged. If you break them out using subqueries you'll have a much easier time making a join. Note that joining like this is not the most efficient method, but then again neither is the given schema...
SELECT Data.*, Limits.Lim1, Limits.Lim2
FROM (SELECT * FROM Files WHERE SubType IS NOT NULL) DataFiles
JOIN (SELECT * FROM Files WHERE SubType IS NULL) LimitFiles
ON LimitFiles.Type = DataFiles.Type
JOIN Data
ON DataFiles.ID = Data.File_ID
JOIN Limits
ON LimitFiles.ID = Limits.File_ID
AND DataFiles.SubType = Limits.SubType
ORDER BY Data.File_ID
UPDATE
To be more specific on how to improve the schema: Currently, the Files table doesn't have a clear way to differentiate between Data and Limit file entries. Aside from this, the Data entries don't have a clear link to a single Limit file entry. Although both of these can be figured out as in the SQL above, such logic might not play well with the query optimizer, and certainly can't guarantee the Data-Limit link that you require.
Consider these options:
Instead of linking to a 'Limit' file via Type, link directly to a Limit entry Id. Set a foreign key on that link to ensure the expected Limit entry is available.
Separate the 'Limit' entries from the 'Data' entries by putting them in a separate table.
Create an index on the foreign key. For that matter, add indices for all foreign keys - SQL Server doesn't do this by default.
Of these, I would consider having a foreign key as essential, and the others as modest improvements.
I hope that someone can help me with my issue. I need to create in a single SELECT statement (the system that we use has some pivot tables in Excel that handle one single SELECT) the following:
I have a INL (Invoice Lines) table, that has a lot of fields, but the important one is the date.
INL_ID DATE
19 2004-03-15 00:00:00.000
20 2004-03-15 00:00:00.000
21 2004-03-15 00:00:00.000
22 2004-03-16 00:00:00.000
23 2004-03-16 00:00:00.000
24 2004-03-16 00:00:00.000
Now, I also have a ILD (Invoice Line Details) that are related by an ID field to the INL table. From the second table I will need to use the scu_qty field to "repeat" values from the first one in my results sheet.
The ILD table values that we need are:
INL_ID scu_qty
19 1
20 1
21 1
22 4
23 4
Now, with the scu_qty I need to repeat the value of the first table and also add one day each record, the scu_qty is the quantity of days of the services that we sell in the ILD table.
So I need to get something like (i'm going to show the INL_ID 22 that you can see has a value different of 1 in the SCU_QTY). The results of the select has to give me something like:
INL_ID DATE
22 2004-03-15 0:00:00
22 2004-03-16 0:00:00
22 2004-03-17 0:00:00
22 2004-03-18 0:00:00
In this information I only wrote the fields that need to be repeated and calculated, of course I will need more fields, but will be repeated from the INL table, so I don't put them so you don't get confused.
I hope that someone can help me with this, it's very important for us this report. Thanks a lot in advance
(Sorry for my English, that isn't my first language)
SELECT INL_ID, scu_qty, CalculatedDATE ...
FROM INL
INNER JOIN ILD ON ...
INNER JOIN SequenceTable ON SequenceTable.seqNo <= ILD.scu_qty
ORDER BY INL_ID, SequenceTable.seqNo
Depending on your SQL flavour you will need to lookup date manipulation functions to do
CalculatedDATE = {INL.DATE + SequenceTable.seqNo (days)}
select INL.INL_ID, `DATE`
from
INL
inner join
ILD on INL.INL_ID = ILD.INL_ID
inner join (
select 1 as qty union select 2 union select 3 union select 4
) s on s.qty <= ILD.scu_qty
order by INL.INL_ID
In instead of that subselect you will need a table if quantity is a bit bigger. Or tell what is your RDBMS and there can be an easier way.