Advance Query with Join - sql

I'm trying to convert a product table that contains all the detail of the product into separate tables in SQL. I've got everything done except for duplicated descriptor details.
The problem I am having all the products have size/color/style/other that many other products contain. I want to only have one size or color descriptor for all the items and reuse the "ID" for all the product which I believe is a Parent key to the Product ID which is a ...Foreign Key. The only problem is that every descriptor would have multiple Foreign Keys assigned to it. So I was thinking on the fly just have it skip figuring out a Foreign Parent key for each descriptor and just check to see if that descriptor exist and if it does use its Key for the descriptor.
Data Table
PI Colo Sz OTHER
1 | Blue | 5 | Vintage
2 | Blue | 6 | Vintage
3 | Blac | 5 | Simple
4 | Blac | 6 | Simple
===================================
Its destination table is this
===================================
DI Description
1 | Blue
2 | Blac
3 | 5
4 | 6
6 | Vintage
7 | Simple
=============================
Select Data.Table
Unique.Data.Table.Colo
Unique.Data.Table.Sz
Unique.Data.Table.Other
=======================================
Then the dual part of the questions after we create all the descriptors how to do a new query and assign the product ID to the descriptors.
PI| DI
1 | 1
1 | 3
1 | 4
2 | 1
2 | 3
2 | 4
By figuring out how to do this I should be able to duplicate this pattern for all 300 + columns in the product. Some of these fields are 60+ characters large so its going to save a ton of space.
Do I use a Array?

Okay, if I understand you correctly, you want all unique attributes converted from columns into rows in a single table (detailstable) that has an id and a description field:
Assuming the schema:
datatable
------------------
PI [PK]
Colo
Sz
OTHER
detailstable
------------------
DI [PK]
Description
You can first get all of the unique attributes into its own table with:
INSERT INTO detailstable (Description)
SELECT
a.description
FROM
(
SELECT DISTINCT Colo AS description
FROM datatable
UNION
SELECT DISTINCT Sz AS description
FROM datatable
UNION
SELECT DISTINCT OTHER AS description
FROM datatable
) a
Then to link up the datatable to the detailstable, I'm assuming you have a cross-reference table defined like:
datadetails
------------------
PI [PK]
DI [PK]
You can then do:
INSERT INTO datadetails (PI, DI)
SELECT
a.PI
b.DI
FROM
datatable a
INNER JOIN
detailstable b ON b.Description IN (a.Colo, a.Sz, a.OTHER)

I reckon you want to split description table for different categories, like - colorDescription, sizeDescription etc.
If that is not practical then I would recommend having an extra column showing an category attribute:
DI Description Category
1 | Blue | Color
2 | Blac | Color
3 | 5 | Size
4 | 6 | Size
6 | Vintage | Other
7 | Simple | Other
And then have primary key in this table as combination of ID and Category column.
This will have less chances for injecting any data errors. It will be also easy to track that down.

Related

Database Design: Inventory with multilevel packaging

I'm trying to develop a database for my inventory. But I have no idea how to keep track of multilevel packaging.
For example:
I currently have a products and positions table
products
Id | Name
================
1013 | Metal
1014 | Wood
positions
id | Name
================
1 | 1-1-1-1
2 | 1-1-1-2
And my inventory table I was thinking of doing something like this:
Let's say I stored 1 box with 1000 Metal and 1 box with 500 Wood at position 1-1-1-1
ItemId | ProductId | Quantity | PositionId
==========================================
1 | 1013 | 1000 | 1
2 | 1014 | 500 | 1
So I'll label those two boxes with a barcode 1 and 2 respectively, so if I scan them, I can check this table to see the product and quantity inside them.
But I can also put these 2 boxes (1 and 2) inside another box (let's call it box 3), which would generate a new barcode for it that, if scanned, will show both previous boxes and its items. And store this box 3 in another position
And I can also put this box 3 inside a pallet, generating a new code and so on. So basically I can multilevel package N times.
What is the best table structure to keep track of all of this? Thanks in advance for any help!
I would add another column to the products table, make it a BIT and maybe call it BOM, BillOfMaterials, or whatever makes sense to you
So your products Table would look like this
Then you could create another table called BillOfMaterials
Quantity is how many of your products are needed to make up your new product. So for this example 2 metal and 1 wood make a pencil.
I was able to make a good structure:
My products and positions are the same but I created a stock table like:
id | product_id | amount | parent_id | position_id
=====================================================
1 | 1013 | 1000 | 4 | 1
2 | 1013 | 1000 | 4 | 1
3 | 1014 | 500 | 4 | 1
4 | 1234 | NULL | NULL | 1
The 1234 (random id) is a box that contains 2000 metal and 500 wood. I dont save this box in the product table.
When I scan the box with id 3, I perform a recursive cte query:
with recursive bom as (
select *, 1 as level
from testing.stock
where id = '4' #scanned id
union all
select c.*, p.level + 1
from testing.stock c
join bom p on c.parent_id = p.id
)
select product_id as product, sum(amount), position_id
from bom b
left join testing.product pd on b.product_id = pd.id
where pd.id is not null
group by product_id, position_id
which returns:
sum | product | position
2000 | 1013 | 3
500 | 1014 | 3
to get by position I just run a variation of the above query. To perform an update I get the Ids inside that box and run a
update testing.stock set position = '2' where id in (#variation of above query)
I hope this helps someone. This works for N packaging level

SQL Select based on condition from another table

I have two tables composing with the following definitions:
A
ID | Name | Description
1 | Bag | It's a bag
2 | Leather | Leather makes the bag
3 | Website | It's a website
4 | Images | Just normal images
5 | Text | Some descriptive text
B
parentID | childID | Quantity
1 | 2 | 5
3 | 4 | 2
3 | 5 | 1
I'm trying to find all items in table A which are of level 0 (meaning they have no parent) and then use that to try and find the items in the following levels.
So far I've gotten to this, which I thought would work, but clearly not...
SELECT *
FROM A
WHERE A.ID in (
SELECT B.parentID
FROM B
WHERE (B.childID is NULL)
);
If anyone could shed some light on this I'd really appreciate it. Also if there's some must-reads for SQL querying please let me know, I'd be glad to learn more about this, as it's something I struggle with.
Thanks in advance.
EDIT:
I've edited the data to hopefully illustrate a little better what I'm trying to achieve. There's several items, some are parents and some are children.
From table B you can tell that to make 1 bag, you need 5 leather (bag is parent of leather with quantity 5). Or that the website is composed of 2 Image and 1 Text (website is parent to both Image quantity 2 and Text quantity 1).
I want to query for the bag and website only, meaning the items in table A which have no parents in table B.
Found it myself, basically what I was looking for was this:
SELECT *
FROM A
WHERE A.ID not in (
SELECT B.childID
FROM B
);
This selects all items from A which have no parents (meaning they are not children of any other item).

Trying to find non-duplicate entries in mostly identical tables(access)

I have 2 different databases. They track different things about inventory. in essence they share 3 common fields. Location, item number and quantity. I've extracted these into 2 tables, with only those fields. Every time I find an answer, it doesn't get all the test cases, just some of the fields.
Items can be in multiple locations, and as a turn each location can have multiple items. The primary key would be location and item number.
I need to flag when an entry doesn't match all three fields.
I've only been able to find queries that match an ID or so, or who's queries are beyond my comprehension. in the below, I'd need a query that would show that rows 1,2, and 5 had issues. I'd run it on each table and have to verify it with a physical inventory.
Please refrain from commenting on it being silly having information in 2 different databases, All I get in response it to deal with it =P
Table A
Location ItemNum | QTY
-------------------------
1a1a | as1001 | 5
1a1b | as1003 | 10
1a1b | as1004 | 2
1a1c | as1005 | 15
1a1d | as1005 | 15
Table B
Location ItemNum | QTY
-------------------------
1a1a | as1001 | 10
1a1d | as1003 | 10
1a1b | as1004 | 2
1a1c | as1005 | 15
1a1e | as1005 | 15
This article seemed to do what I wanted but I couldn't get it to work.
To find entries in Table A that don't have an exactly matching entry in Table B:
select A.*
from A
left join B on A.location = B.location and A.ItemNum = B.ItemNum and A.qty = B.qty
where B.location Is Null
Just swap all the A's and B's to get the list of entries in B with no matching entry in A.

Increasing a +1 to the id without changing the content of a column

I have this random table with random contents.
id | name| mission
1 | aaaa | kitr
2 | bbbb | etre
3 | ccccc| qwqw
4 | dddd | qwert
5 | eeee | potentials
6 | ffffffff | toto
What I want is to add in the above table a column with id=3 with different name and different mission BUT the OLD id =3 I want to have an id = 4 with the name and the mission that it had before when it was id=3, and the OLD id =4 become id=5 with the name and mission of id 5 and so on.
its like i want to enter a column inside of the columns and the below column i want to increase there id +1 but the columns rest the same. example below:
id | name| mission
1 | aaaa | kitr
2 | bbbb | etre
3 | zzzzzz| zzzzz
4 | ccccc| qwqw
5 | dddd | qwert
6 | eeee | potentials
7 | ffffffff | toto
why I want to do this ? I have a table that has 2 CLOB. Inside of those CLOBS there are different queries ex: id =1 has clob of creation of a table id=2 inserts for the columns id=3 has creation of another table id=4 has functions
if you add all of this id in one text(or clob) they will have to create then inserts then create then functions. that table it is like a huge script .
Why I am doing this ? The developers are building their application and they want the sql to work in specific order and I have 6 developers and am organizing the data modeling and the performance and how the scripts are running .So the above table is to organize the calling of the scripts that they wany
Simply put, don't do it.
This case highlights why you should never use any business value, i.e. any 'real world values' for a Primary Key.
In your case I would recommend primary keys not be used for any other purposes.
I recommend you add an extra column 'order' and then change THAT column in order to re-order the rows. That way your primary key and all the other records will not need to be touched.
This avoid the issue that your approach would need to change ALL the database records below the current record which seems like a really bad approach. Just imagine trying to undo that update ;)
Some more info here: https://stackoverflow.com/a/8777574/631619
UPDATE random_table r1
SET id =
(SELECT CASE WHEN id > 2 THEN id+1 ELSE id END id FROM random_table r2
WHERE r1.mission=r2.mission
)
Then insert the new value.

Grouping issue in SSRS

So I have a table that has divisions and sub-divisions. The structure is as follows:
COMPANYID | DIVISIONID | DIVISION | PARENTID
1 | 1 | North | NULL
1 | 2 | South | NULL
1 | 3 | East | NULL
1 | 4 | West | NULL
1 | 5 | Sales | 1
1 | 6 | Service | 1
1 | 7 | Sales | 2
1 | 8 | Service | 2
1 | 9 | Sales | 3
1 | 10 | Service | 3
1 | 11 | Sales | 4
1 | 12 | Service | 4
The idea behind the table is that the data would indicate all of those divisions/subdivisions are the for the same company. Anything with a null PARENTID is a "Division" and anything that has a PARENTID is a "Sub-Division". There will be a max of 2 tiers (Division --> Subdivision). This table would create the following hierarchy:
Company 1
North
Sales
Service
South
Sales
Service
East
Sales
Service
West
Sales
Service
I am trying to make an SSRS report that shows details about the company including a table with all the divisions and sub-divisions. I am using the following query to retrieve a list of all the divisions and subdivisions:
SELECT division, parentid
FROM division
WHERE companyid=#companyid
#companyid is a parameter for the report. My thought process is that I want create a table that has a Parent group and a Detail group where the Parent group would be all those rows that have a null PARENTID and then the Detail group would list all rows whose PARENTID equals the divisionid of Parent row. I'm just not sure how to set that up in the groups. Filters? Separate queries? Do it in the SQL query?
Any and all help is greatly appreciated! Thanks!
For a simple hierarchy with a maximum of 2 levels you can just join the table on itself:
SELECT D1.CompanyId, D1.DivisionId, D1.Division, D2.Division AS Subdivision
FROM DivisionTable D1
INNER JOIN DivisionTable D2 ON D1.DivisionId = D2.ParentId AND D1.CompanyId = D2.CompanyId
WHERE D1.CompanyId = #CompanyId
If i understood your goal correctly, you want the report to display the information in hierarchy mode.
If so you can achieve it with setting a Recursive Hierarchy Group.
Reporting services has the ability to display "hierarchy" information in a "tree" layout
For more information look here
You don't have to create another query the first one that has the CompanyID And the ParentID is just fine
You just need your select statement as a dataset and SSRS can do the grouping for you. You can even tier on multiple levels or better yet collapse and expand your groupings based on filters.
How to get started:
Create your dataset as you have shown
Put in a table object from the toolbox
Only put in the columns of detail data
In the lower left of the 'Design' layout click the 'detail's and click 'add Group>Parent Group'. Select your grouping
If you want you can 'collapse' on demand the details columns. You can choose 'Visibility' on the detail grouping. 'When the report is initially run'>Choose 'Hide'. Check 'Display can be toggled by this report item': Select a cell reference to what you want to expand the details with.