TableView Transform

With the TableView transform, data from a source can be filtered and sorted. Sorting behavior depends on the value type of the sorted column. See more in the Sorting section below.

Any column can be filtered with the following operators. The values for each expression depend on the operator used; specific values are noted when necessary in the descriptions below.

equal Only the specified elements are filtered. Example: Austria.
inAlpharange Alphanumerical values in a particular range are filtered. Examples include:
  • “[A100,A200]” - inclusive range: A100, A101, ... A199, A200
  • “(A100,A200)” - exclusive range: A101, A102, ... A198, A199
  • “[A100,]" - inclusive half-open interval : A100, A101, ...
  • “[,A100)” - exclusive half-open interval: ... A098, A099
inrange Numerical values in a particular range are filtered. Examples include:
  • “[100,200]” - inclusive range: 100, 101.....199, 200
  • “(100,200)” - exclusive range: 101, 102, ... 198, 199
  • “[100,]” - inclusive half-open interval: numbers greater or equal to 100
  • “[,100)” - exclusive half-open interval: numbers lower than 100
isEmpty Empty values (blank, space, or multiple space values). For this operator, the Value field should be left blank.
isNull The condition is true for NULL values but not for (possibly trimmed) empty strings.
like Filters the dimension elements according to regular expressions. General information on regular expressions can be found here.


Examples:

Expression Result
Jan|Feb|Mar Jan, Feb, or Mar
^KS All values starting with KS
_00$ All values ending with _00
^$ Only empty values

Inputname Filter Type Operator Value
Date accept equal 2009

The settings for the TableView transform are described below:

Filter type: various filter lines to the same input column are evaluated like a rule chain with the filter types "accept" and "deny". If only accept-filters are used this behaves like a logical OR.

Logical operator: here you can define the logical operator between the filtering lines. This is only relevant if there are several filter lines with different input columns.

and: a data record is in the filter result when it matches the conditions of all filtered columns.

or: a data record is in the filter result when it matches the conditions of at least one filtered column.

The logical filter operator "and" is the default.

Example:

Data source Source example
Filter type “accept”

Filter example

Result with logical operator “and”

Result example

Result with logical operator “or”

Result example 2

Filter type “deny” Filter example 2
Result with logical operator “and” Result screenshot
Result with logical operator “or” Result screenshot 2

Sorting

The columns of a data source can be sorted. Elements can be sorted in ascending or descending order, with or without case sensitivity. The sorting is processed BEFORE an eventual filtering of start and end rows.

Note that sorting is numerical only when the sorted column has number sequences as value type, e.g. double / integer numbers; otherwise, the default behavior is alphanumeric order, that is, a lexical / string-based sorting. For example:

  • Lexical sorting: 12 < 5 < 5.0 < 6

  • Numerical sorting: 5 < 6 < 12

Filtering of start and end rows

The first and the last rows of the data source can be filtered with these options:

Start line The result starts in the source row with this index. Previous rows are filtered out. If not set, all rows are taken from the beginning.
End line The result ends in the source row with this index. Subsequent rows are filtered out. If not set, all rows are taken up to the end.

Note: when the input is coming dynamically from the Source or Transform, but the "Field name" is left empty, the name of the source column/ function is used for the Field name. When the target is manually defined as constant, then the name "constant" is chosen instead of the value from the "Input" field. Adding more than one constant column results in error, because there would be two columns with the name "constant."

Updated November 4, 2024