SQL Convert Rows to Columns and Columns to Rows without using Pivot Functions
Вставка
- Опубліковано 23 січ 2022
- In this video we will discuss how to convert rows to columns and vice versa without using pivot functions. Below is the script to create table and data:
create table emp_compensation (
emp_id int,
salary_component_type varchar(20),
val int
);
insert into emp_compensation
values (1,'salary',10000),(1,'bonus',5000),(1,'hike_percent',10)
, (2,'salary',15000),(2,'bonus',7000),(2,'hike_percent',8)
, (3,'salary',12000),(3,'bonus',6000),(3,'hike_percent',7);
select * from emp_compensation;
Master the art of SQL with my zero to hero SQL for data analytics course.
www.namastesql.com/course-detail/think-sql-go-from-zero-to-hero-english
Man you might don't know how much you are helping other people. True gem
🙏
Man you are an amazing person helping a lot of people
Thank you for your video. I have an exact requirement and your video makes it very clear with a proper description and makes it easy to understand. Thank you so much
Glad it helped you 😊
Really a great concept, explained very well.
Practising from your videos before my interview. Thanks for the videos.
Best of luck
I was asked this question in an interview (US based Health insurance company) but I only knew how to solve it using Pivot function and could not answer it. I wish if I had found this video sooner!
Great video! Excellent step-by-step explanation. Thank you.
🙏
Thanks Ankit for your time on this video 😊 appreciate all your effort
This was an awesome tutorial! Very informative without the extra puff. Thank you so much!
Glad it was helpful!
this is what I was searching for.. thanks man
I got a query related to this, thank you now i got some clarity.
keep doing what you are doing really it is very helpful.
Very much helpful video. Ankit sir please upload such more complex query video
Sir you are brilliant, so much easier then using SQL pivot operator. Thank you kindly.
Happy to help
@Ankit - Awesome very well explained !
Good one ankit, I am watching daily 1-3 video from your channel to prepare for faang
Awesome
Thank you very much, Sir. The same question I have been asked in the interview. And because I had already watched this video, I could be able to solve this in the interview.
Excellent
Hi ankit, thank you so much for simplifying the concepts. It would be great if you could make similar videos using the sum and case concept solving real interview questions or problems
I have several videos on my channel using that concept. 😊
it's hard to see so so clear and easy understanding video. thank you
Glad it was helpful!
Informative content ❣
Thanks for explanation. It was nice
It's really awesome video. I have learnt some great insights from this video
Glad to hear that!
Your videos has golden content! Just wowwwww!
🙏🙏
I have been working with SQL for a decade now. Now only I know that we can use 'val' for producing result instead of the field name itself. True Genius.
val is a column in the table 😊
@@ankitbansal6 pls suggest how to study from so many videos..is there any pdf of these available.?
@@jececdept.9548 just solve 1 question at a time everyday
Hello ankit sir,
Truly blessed to come across your superb videos....thank you so much sir and please keep helping is this way by more of ur interview questions on sql, pythons
Keep watching
Great work Ankit so helpful thank you.
Glad it helped
Well explained video!
Nice video! Thank you.
This was great and well explained. Thanks
Glad you enjoyed it!
Thanks Ankit for the video, we have to use pivot function using pandas for subsequent ml work, but now we can do the data transformation directly in SQL and post which we can directly train the ml model
Yes, you are right
Thanks a lot sir for explaining in detail.
You are most welcome
Thanks very well explained...!
Glad it was helpful!
Really good explanation, great job :)
Thank you 😊
Thankyou so much for this video
This is very helpful!
Thanks
Glad it was helpful!
Thank you valuable share
Thank you, this help me a lot.
Glad it helped!
So fantastic. Thank you so much. :)
You are so welcome!
True superhero👍
Thank you so much sir for your help. 👍👍🙏
Keep watching 💪
hello sir. thank you for sharing these concepts.its really helpful.sir im an aspiring data analyst and i really want to become good in sql.i know the basics and intermediate concepts.from where should i practice the hard concepts like this which you are teaching?once again thank u.
Thank you for your efforts : )
My pleasure!
Hi Ankit, very crisply explained. Can you explain how each row gets created in both the cases.I mean execution of the query. Thanks in advance!!
Great session 👍
Glad you enjoyed it
very useful 👌👌👍👍
Thank you so much
Very nicely explained
Keep watching😊
🎯 Key Takeaways for quick navigation:
00:14 📊 Learn how to convert rows to columns and vice versa in SQL without using pivot functions.
01:25 🧮 Use SQL case statements to pivot data from rows to columns for specific components.
04:27 🧾 Practice using case statements with sum for efficient data pivoting in SQL.
05:28 🔄 Unpivot data by creating a new table using SELECT INTO and then using UNION to achieve the desired format.
08:18 📚 Mastering case when with sum is crucial for handling data transformations, and it's a valuable skill for interviews and job tasks.
Made with HARPA AI
Thanks friend very helpful
Welcome
Great video
Glad you enjoyed it
Thank you so much ! 🙂
You're welcome 😊
thank you so much
super sir make a more interview quations.....
i have also tried this approach using window functions -
select emp_id, salary, hike_percent, bonus
from (
select emp_id, val as salary,
lead(val) over a as hike_percent,
lead(val,2) over a as bonus,
row_number() over a as rn
from emp_compensation
window a as (partition by emp_id order by salary_component_type desc)
) t
where rn=1
Thank you
Well explained... Please make some videos on path to join Amazon as a Business Intelligence Engineer or Data Engineer.
Sure.
Thanks for sharing much value , a small request if you can make such more videos
god bless you
Sure 😊
@@ankitbansal6 People like you are a great help to the society and for new learners, we pray many people support this channel
Will inform other people also to join this channel
Hello Ankit,
I bought multiple courses on SQL but finally started to understand by watching your videos. Thank you so much.
I want to know where can I get more use cases to practice for case statements using sum and other aggregate functions.
Check out the complex SQL problems playlist
Hi Ankit, Thanks for creating this video . It was really helpful . Can you please create more examples around this.
Sure
Hi Ankit
I am learning a lot from you.
Thank you, Thank you again.
Please clear my doubt.
Pasting my query below.
I tried this method.
but following both this method and aggregating both ways coming same output
select
emp_id,
sum(case when salary_component_type = 'salary' then val end) as salary,
sum(case when salary_component_type = 'bonus' then val end) as bonus,
sum(case when salary_component_type = 'hike_percent' then val end) as hike_percent
from emp_compensation
group by emp_id;
select
emp_id,
max(case when salary_component_type = 'salary' then val end) as salary,
max(case when salary_component_type = 'bonus' then val end) as bonus,
max(case when salary_component_type = 'hike_percent' then val end) as hike_percent
from emp_compensation
group by emp_id;
how???
There is only one not null value in each group. So max min avg all same ..
@@ankitbansal6 ok
excellent
Hi Ankit, if we have other salary component we have to alter the query . Can we dynamically generate the columns(for each Salary component)without using the pivot function.
Nice Video, thank you. I have a doubt. Is there any way that we can dynamically create these columns during pivoting instead of explicitly typing the column values in each sum(case) statement? If we have more unique values and more rows, it might be useful if possible.
There is pivot function to do that. I will make a separate video.
I am a big fan if you Ankit 💓
Thank you 😊
Thanks a lot Baha’i, very helpful video. I tried this in report which has 200 columns to unpivot and 5 inner joins, finally the script has 2300 lines now (each union has 15 lines). I have used inner joins on each union. Is it the a way to refine the script more!
Hello Ankit, firstly congratulations for hitting 1lakh subscribers on youtube and gratitude for the resourceful videos. I have one doubt while trying to unpivot the table i am encountering the error msg Undeclared variable : emp_compensation_pivot , how to remove that error. Pls help. Thanks in advance.
Hi Ankit, Your videos are helpful to learn SQL in depth, do you have any github repository where you have hosted all your YT questions and solutions ? Or something in text/ blog format ?
I haven't done that yet. Planning that.
Very well explained. I just have one question which may sound very basic. How to distinguish between pivoting and unpivoting? i always get confused which one is what
Row to column is pivot and reverse in unpivot.
Thank you Ankit for awesome explanation.
Is there another video to practice this topic?
There are many in this playlist
Complex SQL Questions for Interview Preparation: ua-cam.com/play/PLBTZqjSKn0IeKBQDjLmzisazhqQy4iGkb.html
Thank you . What if I add one more type called Gender that has "Male" or "Female" - we cannot do SUM or any numerical aggregate like MIN/MAX in that case - so how do we solve that?
awesome vide
Glad you enjoyed it
Thanks a lot Ankit it was very helpful. Just one question in case we have multiple column like let's say 10-15 then we need to write 10-15 case statement from row to column and 10-15 union all in case od column to row or do we have any other approach
Alternatively you can use pivot and unpivot functions.
Well explained , just a request Please make a video on how to upload DB tables on this Software , it would be a great help from ur end !!
You mean you want to load data from files in your system to SQL server right ?
@@ankitbansal6 YES
Please make a video on using Pivot functions and merge statement
Sure.
Is 'select into' syntax for creating new table worked in MySQL ? I got error
hi bro actually i tried the same but while doing unpivoting .... "select emp_id, 'salary' as salary_component_type, salary as val from temp" i got an error code 1054: unknown column salary in field list (where i actually want 10000 as output in val column). can u help me with this.
Incase if the rows are way to high how do we dynamically bring them as column names, is there a way to do it? It's not possible to give names individually . In my case I should bring dates as column names from the row.
This solved my potential career-ending problem, thanks a lot brother. How can I perform high performance on this when I am working on big data?
It's optimised technique only.
Hi Ankit, if possible pls create video on how to do pivoting dynamically. If new entries comes in future it will automatically comes from rows to columns
Okay
Hi Ankit great video. Just one doubt can sum function work on string values in the case statement? Example just adding another columns as name and pan no.
Yes it works on string.
@@ankitbansal6 thank you for the clarification. 😊
How to pivot If all my columns having date. If sum will work in that case?
Hi i have try this example in DB but this is not working it show me error when i use case statement and using then val showing invalid column name please suggest me.
The same query how can we perform if the value has non numeric values.
I could not understand sir we can do it by pivot function right ? can u please explain ?
This method is dependent on number of columns. How to make this dynamic like pivot. Pivot automatically takes new column into account whereas this method will require to change query to add every new column. Please help.
Hi Ankit, i am trying to run the below query in MySQL but getting error like undeclared variable at into emp_compensation_pivot
Select emp_id
,SUM(case when salary_component_type = 'salary' then val end) as Salary
,SUM(case WHEN salary_component_type = 'bonus' then val end) as Bonus
,SUM(case when salary_component_type = 'hike_percent' then val end) as hike_percent
into emp_compensation_pivot
from emp_compensation
group by emp_id;
Can you please guide me
Use create table as instead of into
@@ankitbansal6
sir used create table statement also but it is not working
is their any other alternative?
What if there are 2 columns example val and expense_amount. So i want 2 colums from 1 row. Salary_val and salary_expense_amount. How to do that?
Hello Ankit! This was really helpful, thank you so much for explaining it step by step. I came across a similar problem but have quite a few doubts, could you please help me understand how the same concept can be applied for the below table:
NAME PROFESSION
Samantha Doctor
Julia Actor
Maria Actor
Meera Singer
Ashley Professor
Ketty Professor
Christeen Professor
Jane Actor
Jenny Doctor
Priya Singer
Here is the query that I used:
select
case when profession = 'doctor' then person_name end as 'doctor',
case when profession = 'actor' then person_name end as 'actor',
case when profession = 'professor' then person_name end as 'professor',
case when profession = 'singer' then person_name end as 'singer'
from occupation
How do I exclude the nulls is this case?
You just need to create one case statement with multiple when ..that will do
Hi Hiya, Please post your solution . It will be helpful
I guess this is a hackerrank question, you can just group by and take min/max.
Hello Hiya
I tried to solve the same problem using the above mentioned method. The above example has emp-id, we can apply GROUP BY on this emp-id and get the desired result(remove the null values).
However in the problem, we dont have any such column to perform GROUP BY on and eliminate the nulls.
To overcome this we create a new column ‘grouper’. What grouper does is, it ranks the candidates for each profession.
For ex: If there are 4 doctors, we give a rank to each doctor. Similarly for other professions.
Then we perform GROUP BY on grouper column and get the desired result.
Here’s the piece of query I have used.
SET @r1=0, @r2=0, @r3=0, @r4=0;
WITH tb1 AS
(SELECT
CASE
WHEN Occupation = 'Doctor' THEN Name
END AS Doctor,
CASE
WHEN Occupation = 'Professor' THEN Name
END AS Professor,
CASE
WHEN Occupation = 'Singer' THEN Name
END AS Singer,
CASE
WHEN Occupation = 'Actor' THEN Name
END AS Actor,
CASE
WHEN Occupation = 'Doctor' THEN (@r1:=@r1+1)
WHEN Occupation = 'Professor' THEN (@r2:=@r2+1)
WHEN Occupation = 'Singer' THEN (@r3:=@r3+1)
WHEN Occupation = 'Actor' THEN (@r4:=@r4+1)
END AS grouper
FROM OCCUPATIONS
ORDER BY Name)
SELECT
MIN(Doctor) AS Doctor,
MIN(Professor) AS Professor,
MIN(Singer) AS Singer,
MIN(Actor) AS Actor
FROM tb1
GROUP BY grouper;
Hope this helps !
Thanks for your video, but what if I want to unpivot 35 different months columns into row based and 35 quantity columns for the corresponding month into row based.
This method is not useful fod that, also the unpivot function is not efficient here, query processing times is going more then an hour, could you please suggest any approach for this
You have only 2 options. Either use pivot unpivot functions or do like in video.
what happend when someone enter the new salary component type
the query set was only the exitsting data when someone enter the new salary component type
so every time i change the query
it's not good
please help me this
like whenever some add new salary component type how to automatically add val in query
You need to modify the query. Or wrote a dynamic SQL using pl SQL or t SQL
Hi Ankit, Nice video, thanks for this one. Is there a video, to convert rows to columns with column name dynamically, instead of explicit mention in the case statement and pivot function.
Note: Sybase doesn't have Pivot function.
Sure will look into it 🙂
Thank you! Looking forward to your video on it
Please make video on indexing
Sure. That's in my list.
This is awesome!! Would this syntax work in mysql too?
Yes it will
Sir make please plsql series.
This is not dynamic. This will only work when we know data fully.
can this process be used another query where numbers are not involved for pivot
Hi Ankit, This was very informative. I had the a bit similar usecase in my work but the second column will have 1000s uniques and third columns have respective values. I was able to solve it using Python but the transformation was slow. Is there a way to extend this solution with or without PIVOT for 1000s of unique records ?
Which means you will have 1000s of column in your o/p?
@@ankitbansal6 Correct. In short we need to produce the exact same logic for large dataset. how to scale up this solution for 1000 unique values of second column of the original dataset ?
@@rakeshchaudhary8255 there is a pivot function. I will make a video on it.
@@ankitbansal6 Will search for this .I would like to understand how this can be implemented.
Does any one know how to solve this problem if we have strings instead of integers. I can't add string right what is the alternative?
Hi Ankit, Really love your content, my SQL skills are getting better because of this series. Thanks a lot. Can you please tell a good source to practice SQL online with questions of these standards
Good to know that 😊 I think if you solve all my videos. It is good enough 🙂
@@ankitbansal6 Thanks, will do that
Hi Ankit, thank you sharing such a wonderful tip. this solves purpose.. But i have text value in "Val" column. when i use sum function, it shows error invalid digit value 'L'. any idea to group text values
You can use string agg function ..or can use min , max