MySQL IntroductionComment MySQLInstall MySQL container with Docker-ComposeMySQL AdminMySQL ALTER TABLEMySQL ArithmeticMySQL BackticksMySQL Backup using mysqldumpMySQL Change PasswordMySQL Character Sets and CollationsMySQL clientMySQL ClusteringMySQL Configuration and tuningMySQL Connecting with UTF-8 Using Various Programming language.MySQL Converting from MyISAM to InnoDBMySQL Create New UserMySQL Creating databasesMySQL Customize PS1MySQL Data TypesMySQL Date and Time OperationsMySQL Dealing with sparse or missing dataMySQL DELETEMySQL Drop TableMySQL Dynamic Un-Pivot Table using Prepared StatementMySQL ENUMMySQL Error 1055: ONLY FULL GROUP BY: something is not in GROUP by clause ...MySQL Error codesMySQL EventsMySQL Extract values from JSON typeMySQL Full-Text searchMySQL Group byMySQL Handling Time ZonesMySQL Indexes and KeysMySQL INSERTMySQL JoinsMySQL JOINS: Join 3 table with the same name of id.MySQL JSONMySQL Limit and OffsetMySQL LOAD DATA INFILEMySQL LOCK TABLEMySQL Log filesMySQL Many-to-many Mapping tableMySQL MyISAM EngineMySQL mysqlimportMySQL NULLMySQL One to ManyMySQL ORDER byMySQL PartitioningMySQL Performance TipsMySQL Performance TuningMySQL Pivot queriesMySQL PREPARE StatementsMySQL Recover from lost root passwordMySQL Regular ExpressionsMySQL ReplicationMySQL Reserved WordsMySQL Security via GRANTsMySQL SELECTMySQL Server InformationMySQL SSL Connection SetupMySQL Stored routines (procedures and functions)MySQL String operationsMySQL Table CreationMySQL Temporary TablesMySQL Time with subsecond precisionMySQL TransactionMySQL TRIGGERSMySQL UNIONMySQL UnionsMySQL UPDATEMySQL Using VariablesMySQL VIEWRecover and reset the default root password for MySQL 5.7+

MySQL Dynamic Un-Pivot Table using Prepared Statement

From WikiOD

Un-pivot a dynamic set of columns based on condition[edit | edit source]

The following example is a very useful basis when you are trying to convert transaction data to un-pivoted data for BI/reporting reasons, where the dimensions which are to be un-pivoted can have a dynamic set of columns.

For our example, we suppose that the raw data table contains employee assessment data in the form of marked questions.

The raw data table is the following:

create table rawdata

 PersonId VARCHAR(255)
,Question1Id INT(11)
,Question2Id INT(11)
,Question3Id INT(11)

The rawdata table is a temporary table as part of the ETL procedure and can have a varying number of questions. The goal is to use the same un-pivoting procedure for an arbitrary number of Questions, namely columns that are going to be un-pivoted.

Below is a toy example of rawdata table:


The well-known,static way to unpivot the data, in MYSQL is by using UNION ALL:

create table unpivoteddata


 PersonId VARCHAR(255)
,QuestionId VARCHAR(255)
,QuestionValue INT(11)


INSERT INTO unpivoteddata SELECT PersonId, 'Question1Id' col, Question1Id 
FROM rawdata
SELECT PersonId, 'Question2Id' col, Question2Id 
FROM rawdata
SELECT PersonId, 'Question3Id' col, Question3Id 
FROM rawdata;

In our case we want to define a way to unpivot an arbitrary number of QuestionId columns. For that we need to execute a prepared statement that is a dynamic select of the desired columns. In order to be able to choose which columns need to be un-pivoted, we will use a GROUP_CONCAT statement and we will choose the columns for which the data type is set to 'int'. In the GROUP_CONCAT we also include all additional elements of our SELECT statement to-be executed.

set @temp2 = null; 

SELECT GROUP_CONCAT(' SELECT ', 'PersonId',',','''',COLUMN_NAME,'''', ' col     ',',',COLUMN_NAME,' FROM rawdata' separator ' UNION ALL' ) FROM     INFORMATION_SCHEMA.COLUMNS WHERE table_name = 'rawdata' AND DATA_TYPE = 'Int' INTO     @temp2;

select @temp2;

In another occasion we could have chosen columns that the column name matches a pattern, for example instead of




or something suitable that can be controlled through the ETL phase.

The prepared statement is finalized as follows:

set @temp3 = null;

select concat('INSERT INTO unpivoteddata',@temp2) INTO @temp3;

select @temp3;

prepare stmt FROM @temp3;
execute stmt;
deallocate prepare stmt;

The unpivoteddata table is the following:

SELECT * FROM unpivoteddata


Selecting columns according to a condition and then crafting a prepared statement is an efficient way of dynamically un-pivoting data.