postgres: Multiply column of table A with rows of table B - sql

Fellow SOers,
Currently I am stuck with the following Problem.
Say we have table "data" and table "factor"
"data":
---------------------
| col1 | col2 |
----------------------
| foo | 2 |
| bar | 3 |
----------------------
and table "factor" (the amount of rows is variable)
---------------------
| name | val |
---------------------
| f1 | 7 |
| f2 | 8 |
| f3 | 9 |
| ... | ... |
---------------------
and the following result should look like this:
---------------------------------
| col1 | f1 | f2 | f3 | ...|
---------------------------------
| foo | 14 | 16 | 18 | ...|
| bar | 21 | 24 | 27 | ...|
---------------------------------
So basically I want the column "col2" multiplicated with all the contents of "val" of table "factor" AND the content of column "name" should act as tableheader/columnname for the result.
We are using postgres 9.3 (upgrade to higher version may be possible), so an extended Search resulted in multiple possible solutions: using crosstab (though even with crosstab I was not able to figure this one out), using CTE "With" (preferred, but also no luck). Probably this may also be done with the correct use of array() and unnest().
Hence, any help is appreciated on how to achieve this (the less code, the better)
Tnx in advance!

This package seems to do what you want:
https://github.com/hnsl/colpivot

Related

Select data from multiple existing tables dynamically

I have tables "T1" in the database that are broken down by month of the form (table_082020, table_092020, table_102020). Each contains several million records.
+----+----------+-------+
| id | date | value |
+----+----------+-------+
| 1 | 20200816 | abc |
+----+----------+-------+
| 2 | 20200817 | xyz |
+----+----------+-------+
+----+----------+-------+
| id | date | value |
+----+----------+-------+
| 1 | 20200901 | cba |
+----+----------+-------+
| 2 | 20200901 | zyx |
+----+----------+-------+
There is a second table "T2" that stores a reference to the primary key of the first one and actually to the table itself only without the word "table_".
+------------+--------+--------+--------+--------+
| rec_number | period | field1 | field2 | field3 |
+------------+--------+--------+--------+--------+
| 777 | 092020 | aaa | bbb | ccc |
+------------+--------+--------+--------+--------+
| 987 | 102020 | eee | fff | ggg |
+------------+--------+--------+--------+--------+
| 123456 | 082020 | xxx | yyy | zzz |
+------------+--------+--------+--------+--------+
There is also a third table "T3", which is the ratio of the period and the table name.
+--------+--------------+
| period | table_name |
+--------+--------------+
| 082020 | table_082020 |
+--------+--------------+
| 092020 | table_092020 |
+--------+--------------+
| 102020 | table_102020 |
+--------+--------------+
Tell me how you can combine 3 tables to get dynamic data for several periods. For example: from 15082020 to 04092020, where the data will be located in different tables, respectively
There really is no good reason for storing data in this format. It makes querying a nightmare.
If you cannot change the data format, then add a view each month that combines the data:
create view t as
select '202010' as YYYYMM, t.*
from table_102020
union all
select '202008' as YYYYMM, t.*
from table_092020
union all
. . .;
For a once-a-month effort, you can spend 10 minutes writing the code and do so with a calendar reminder. Or, better yet, set up a job that uses dynamic SQL to generate the code and run this as a job after the underlying tables are using.
What should you be doing? Well, 5 million rows a months isn't actually that much data. But if you are concerned about it, you can use table partitioning to store the data by month. This can be a little tricky; for instance, the primary key needs to include the partitioning key.

Primary key auto-increment manipulation

Is there any way to have a primary key with a feature that increments it but fills in gaps? Assuming I have the following table:
____________________
| ID | Value |
| 1 | A |
| 2 | B |
| 3 | C |
^^^^^^^^^^^^^^^^^^^^^
Notice that the value is only an example, the order has nothing to do with the question.
Once I remove the row with the ID of 2 (the table will look like this):
____________________
| ID | Value |
| 1 | A |
| 3 | C |
^^^^^^^^^^^^^^^^^^^^^
And I add another row, with regular auto-increment feature it will look like this:
____________________
| ID | Value |
| 1 | A |
| 3 | C |
| 4 | D |
^^^^^^^^^^^^^^^^^^^^^
As expected.
The output I'd want would be:
____________________
| ID | Value |
| 1 | A |
| 2 | D |
| 3 | C |
^^^^^^^^^^^^^^^^^^^^^
Where the gap is filled with the new row. Also note that maybe, in memory, it would look different. But the point is that the primary key would fill the gaps.
When having the primary keys (for instance) 1, 2, 3, 6, 7, 10, 11, 4 should be first filled in, then 5, 8 and so on... When the table is empty (even if it had a million of rows before) it should start over from 1.
How do I accomplish that? Is there any built-in feature similar to that? Can I implement it?
EDIT: If it's not possible, why not?
No, you don't want to do that, as juergen-d said. It's unlikely to do what you think it is doing, and it will do it even less in a multi-user environment.
In a multiuser environment you are likely to get voids even when there are no deletes, just from aborted inserts.

Unique string table in SQL and replacing index values with string values during query

I'm working on an old SQL Server database that has several tables that look like the following:
|-------------|-----------|-------|------------|------------|-----|
| MachineName | AlarmName | Event | AlarmValue | SampleTime | ... |
|-------------|-----------|-------|------------|------------|-----|
| 3 | 180 | 8 | 6.780 | 2014-02-24 | |
| 9 | 67 | 8 | 1.45 | 2014-02-25 | |
| ... | | | | | |
|-------------|-----------|-------|------------|------------|-----|
There is a separate table in the database that only contains unique strings, as well as the index for each unique string. The unique string table looks like this:
|----------|--------------------------------|
| Id | String |
|----------|--------------------------------|
| 3 | MyMachine |
| ... | |
| 8 | High CPU Usage |
| ... | |
| 67 | 404 Error |
| ... | |
|----------|--------------------------------|
Thus, when we want to get something out of the database, we get the respective rows out, then lookup each missing string based on the index value.
What I'm hoping to do is to replace all of the string indexes with the actual values in a single query without having to do post-processing on the query result.
However, I can't figure out how to do this in a single query. Do I need to use multiple JOINs? I've only been able to figure out how to replace a single value by doing something like -
SELECT UniqueString.String AS "MachineName" FROM UniqueString
JOIN Alarm ON Alarm.MachineName = UniqueString.Id
Any help would be much appreciated!
Yes, you can do multiple joins to the UniqueStrings table, but change the order to start with the table you are reporting on and use unique aliases for the joined table. Something like:
SELECT MN.String AS 'MachineName', AN.String as 'AlarmName' FROM Alarm A
JOIN UniqueString MN ON A.MachineName = MN.Id
JOIN UniqueString AN ON A.AlarmName = AN.Id
etc for any other columns

Grouped string aggregation / LISTAGG for SQL Server

I'm sure this has been asked but I can't quite find the right search terms.
Given a schema like this:
| CarMakeID | CarMake
------------------------
| 1 | SuperCars
| 2 | MehCars
| CarMakeID | CarModelID | CarModel
-----------------------------------------
| 1 | 1 | Zoom
| 2 | 1 | Wow
| 3 | 1 | Awesome
| 4 | 2 | Mediocrity
| 5 | 2 | YoureSettling
I want to produce a dataset like this:
| CarMakeID | CarMake | CarModels
---------------------------------------------
| 1 | SuperCars | Zoom, Wow, Awesome
| 2 | MehCars | Mediocrity, YoureSettling
What do I do in place of 'AGG' for strings in SQL Server in the following style query?
SELECT *,
(SELECT AGG(CarModel)
FROM CarModels model
WHERE model.CarMakeID = make.CarMakeID
GROUP BY make.CarMakeID) as CarMakes
FROM CarMakes make
http://www.simple-talk.com/sql/t-sql-programming/concatenating-row-values-in-transact-sql/
It is an interesting problem in Transact SQL, for which there are a number of solutions and considerable debate. How do you go about producing a summary result in which a distinguishing column from each row in each particular category is listed in a 'aggregate' column? A simple, and intuitive way of displaying data is surprisingly difficult to achieve. Anith Sen gives a summary of different ways, and offers words of caution over the one you choose...
If it is SQL Server 2017 or SQL Server VNext, Azure SQL database you can use String_agg as below:
SELECT make.CarMakeId, make.CarMake,
CarModels = string_agg(model.CarModel, ', ')
FROM CarModels model
INNER JOIN CarMakes make
ON model.CarMakeId = make.CarMakeId
GROUP BY make.CarMakeId, make.CarMake
Output:
+-----------+-----------+---------------------------+
| CarMakeId | CarMake | CarModels |
+-----------+-----------+---------------------------+
| 1 | SuperCars | Zoom, Wow, Awesome |
| 2 | MehCars | Mediocrity, YoureSettling |
+-----------+-----------+---------------------------+

Eliminate full table scan due to BETWEEN (and GROUP BY)

Description
According to the explain command, there is a range that is causing a query to perform a full table scan (160k rows). How do I keep the range condition and reduce the scanning? I expect the culprit to be:
Y.YEAR BETWEEN 1900 AND 2009 AND
Code
Here is the code that has the range condition (the STATION_DISTRICT is likely superfluous).
SELECT
COUNT(1) as MEASUREMENTS,
AVG(D.AMOUNT) as AMOUNT,
Y.YEAR as YEAR,
MAKEDATE(Y.YEAR,1) as AMOUNT_DATE
FROM
CITY C,
STATION S,
STATION_DISTRICT SD,
YEAR_REF Y FORCE INDEX(YEAR_IDX),
MONTH_REF M,
DAILY D
WHERE
-- For a specific city ...
--
C.ID = 10663 AND
-- Find all the stations within a specific unit radius ...
--
6371.009 *
SQRT(
POW(RADIANS(C.LATITUDE_DECIMAL - S.LATITUDE_DECIMAL), 2) +
(COS(RADIANS(C.LATITUDE_DECIMAL + S.LATITUDE_DECIMAL) / 2) *
POW(RADIANS(C.LONGITUDE_DECIMAL - S.LONGITUDE_DECIMAL), 2)) ) <= 50 AND
-- Get the station district identification for the matching station.
--
S.STATION_DISTRICT_ID = SD.ID AND
-- Gather all known years for that station ...
--
Y.STATION_DISTRICT_ID = SD.ID AND
-- The data before 1900 is shaky; insufficient after 2009.
--
Y.YEAR BETWEEN 1900 AND 2009 AND
-- Filtered by all known months ...
--
M.YEAR_REF_ID = Y.ID AND
-- Whittled down by category ...
--
M.CATEGORY_ID = '003' AND
-- Into the valid daily climate data.
--
M.ID = D.MONTH_REF_ID AND
D.DAILY_FLAG_ID <> 'M'
GROUP BY
Y.YEAR
Update
The SQL is performing a full table scan, which results in MySQL performing a "copy to tmp table", as shown here:
+----+-------------+-------+--------+-----------------------------------+--------------+---------+-------------------------------+--------+-------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+-------+--------+-----------------------------------+--------------+---------+-------------------------------+--------+-------------+
| 1 | SIMPLE | C | const | PRIMARY | PRIMARY | 4 | const | 1 | |
| 1 | SIMPLE | Y | range | YEAR_IDX | YEAR_IDX | 4 | NULL | 160422 | Using where |
| 1 | SIMPLE | SD | eq_ref | PRIMARY | PRIMARY | 4 | climate.Y.STATION_DISTRICT_ID | 1 | Using index |
| 1 | SIMPLE | S | eq_ref | PRIMARY | PRIMARY | 4 | climate.SD.ID | 1 | Using where |
| 1 | SIMPLE | M | ref | PRIMARY,YEAR_REF_IDX,CATEGORY_IDX | YEAR_REF_IDX | 8 | climate.Y.ID | 54 | Using where |
| 1 | SIMPLE | D | ref | INDEX | INDEX | 8 | climate.M.ID | 11 | Using where |
+----+-------------+-------+--------+-----------------------------------+--------------+---------+-------------------------------+--------+-------------+
Answer
After using the STRAIGHT_JOIN:
+----+-------------+-------+--------+-----------------------------------+---------------+---------+-------------------------------+------+---------------------------------+
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
+----+-------------+-------+--------+-----------------------------------+---------------+---------+-------------------------------+------+---------------------------------+
| 1 | SIMPLE | C | const | PRIMARY | PRIMARY | 4 | const | 1 | Using temporary; Using filesort |
| 1 | SIMPLE | S | ALL | PRIMARY | NULL | NULL | NULL | 7795 | Using where |
| 1 | SIMPLE | SD | eq_ref | PRIMARY | PRIMARY | 4 | climate.S.STATION_DISTRICT_ID | 1 | Using index |
| 1 | SIMPLE | Y | ref | PRIMARY,STAT_YEAR_IDX | STAT_YEAR_IDX | 4 | climate.S.STATION_DISTRICT_ID | 1650 | Using where |
| 1 | SIMPLE | M | ref | PRIMARY,YEAR_REF_IDX,CATEGORY_IDX | YEAR_REF_IDX | 8 | climate.Y.ID | 54 | Using where |
| 1 | SIMPLE | D | ref | INDEX | INDEX | 8 | climate.M.ID | 11 | Using where |
+----+-------------+-------+--------+-----------------------------------+---------------+---------+-------------------------------+------+---------------------------------+
Related
http://dev.mysql.com/doc/refman/5.0/en/how-to-avoid-table-scan.html
http://dev.mysql.com/doc/refman/5.0/en/where-optimizations.html
Optimize SQL that uses between clause
Thank you!
ONE Request... It looks like you KNOW your data. Add the keyword "STRAIGHT_JOIN" and see the results...
SELECT STRAIGHT_JOIN ... the rest of your query...
Straight-join tells MySql to DO IT AS I HAVE LISTED. So, your CITY table is the first in the FROM list, thus indicating you expect that to be your primary... Additionally, your WHERE clause of the CITY is the immediate filter. With that being said, it will probably fly through the rest of the query...
Hope it helps... Its worked for me with gov't data of millions of records queried and joined to 10+ lookup tables where mySql was trying to think for me.
in order to do efficient between queries you are going to want a b tree index on your YEAR column. for example:
CREATE INDEX id_index USING BTREE ON YEAR_REF (YEAR);
BTREE indexes allow for efficient range queries, if this is in fact the root problem then having an index like this should get rid of the full table scan and have it only scan the part of the table that is in the range. read more about btrees on wikipedia
However, as with any optimisation advice, you should measure to make sure that you don't do more harm than good.
Can you change from searching within a radius to search in a bounding box?
You know the city so you can calculate a bounding box in your application.
Perhaps this
S.LATITUDE_DECIMAL >= latitude_lower and
S.LATITUDE_DECIMAL <= latitude_upper and
S.LONGITUDE_DECIMAL >= longitude_lower and
S.LONGITUDE_DECIMAL <= longitude_upper
could be a little faster?