This article describes the transforms you can use when creating a data cube (ETL process).
The Tabular Select transform is created when you drag a structure onto the data cube designer canvas from a data connector that supports tabular data (e.g. XML, CSV, SQL Server table-valued functions).
The SQL Select transform is created when a structure is dragged onto the canvas from a data connector that supports relational queries.
The MDX Select transform is created when a SSAS cube is dragged onto the canvas from the Data Connectors structure.
Enter an SQL statement to make a selection from a data connector instead of dragging native structures onto the canvas.
The Stored Procedure Select transform lets you retrieve data using a relational stored procedure.
The Aggregate transform allows use of aggregate functions such as SUM, AVG, COUNT, MIN, MAX, etc.
The Calculated Element transform lets you create new elements by writing DundasScript expressions.
The Data Conversion transform allows change of data type of a column to another data type.
The Filter transform filters out rows that do not meet the configured criteria/settings.
The Join transform allows joining two tables by defining the keys and specifying the join type. If relationship exists between the two tables, the link is automatically created, but can be changed if necessary.
The Lookup transform joins data from input columns with column in a lookup table.
The String transform manipulates string columns in data tables by applying string functions.
The Union transform combines data from multiple structures by mapping columns onto one another.
The Math transform lets you perform simple math functions such as Absolute, Ceiling, and Square Root on numeric input columns.
The Copy Element transform creates new columns by copying selected input columns and adding the new columns to the output.
The Fuzzy Grouping transform allows grouping of records by looking at the similarity between the values of various columns.
The Fuzzy Lookup transform allows searching for a matching recording from a secondary table when no relationship key fields are defined between two tables.
By specifying a rate, this transform reads in all of the data from the previous transform and generates a set of random indexes according to the rate input multiplied by the total record count.
The Record Sampling transform reads in all of the data from the previous transform and generates a set of random indexes according to the number input. Then output the records according to those indexes.