In this article, you will find information and examples on the following topics associated with working with JSON and XML data sources.
- JSON and XML data structure
- Selecting a JSON or XML data source
- Using the Selection options in JSON and XML data
- XPath Expressions (with examples)
- XPath Axes
- Klipfolio XPath Functions for JSON data
- Advanced Usage
- JSON and XML data sources: Videos
About the structure of JSON and XML data
JSON and XML are hierarchical data formats.
Where tabular data is referenced by column and row, JSON and XML data is referenced by its path, which indicates its position in the hierarchy. This path, also called the XPath, describes the levels traversed to reach the elements. For example, in the following data source:
player/first_nameis the XPath of the
player/team/whais the XPath of the
JSON data is made up of these elements:
- objects: one or more field name and value pairs, indicated in Klipfolio by curly braces.
- arrays: lists of elements of the same type, indicated in Klipfolio by square brackets.
Selecting a JSON or XML data source
The first image below shows how to choose a data source in the formula editor. The second image shows how the XPath is displayed in the formula once the data source is chosen.
outlined in purple are the data source reference and enclose the editable XPath, which is outlined in red.
is only displayed when the XPath is being edited.
Using the Selection Options in JSON or XML data
When you select data from a JSON or XML data source, the selected field’s path, or XPath, is displayed in the formula editor. You can type the path or you can click on the Selection Options link.
In the following image, the XPath of the
first_name element is
By default, the values of all peer elements are selected (equivalent to selecting a column of data in a table). If you need to change this selection, click Selection Options, and select the required option:
- Select only this <field name> element. Only one specific <field_name> element is selected (equivalent to selecting a specific cell in Excel). The formula editor displays the specific path:
- Select all peer <field name> elements. All <field_name> elements at this level of the hierarchy are selected (equivalent to selecting a column of data in a table). The formula editor displays the path:
- Select all <field name> elements in all <parent> elements, adding blanks for missing <field name> elements. All <field_name> elements in all parents are selected, and blanks are inserted for missing <field_name> elements. The formula editor displays the path:
- Select all field names in <parent> element. All field names in this branch of the hierarchy are selected. The formula editor displays the path:
@ kf:names(player). Learn more here.
- Select all <field_name> elements. All <field_name> elements at any level of the hierarchy are selected. The formula editor displays the path:
- Select all <field_name> elements in this <parent> element. All <field_name> elements in this branch of the hierarchy are selected. The formula editor displays the path:
- Select everything in <parent> elements. Everything (all values) in all <parent> elements is selected. The formula editor displays the path:
An XPath expression is used to select elements based on specified criteria.
Some common functions used in XPath expressions:
Example 1: select elements by matching a specific value
To select the last names of players on team Winnipeg:
- Select the last_name of a player:
@ /player/last_nameThis will return all
- Type the expression,
/playerto select only the last names of players on team Winnipeg.
The resulting XPath:
This says: for all player elements with descendant team/name equal to Winnipeg, select the last_name element.
Example 2: select elements based on a field name
This example uses the contains function and the name function to select all elements based on a field name.
This says: for all /player elements, select any values where the field name contains: "name"
This expression returns Anders, Hedberg, Maurice, Richard, Rocket, Wayne, Gretzky, The Great One.
Example 3: select all descendant values
selects all immediate descendant values in the player element.
The Great One
selects all immediate descendant values (denoted by /*) in the second player element.
XPath axes (child, descendant, following-sibling, preceding-sibling) indicate the position of an element relative to another element in the hierarchy. For example,
- first_name, last_name, nickname, teams/name, teams/wha and teams/nhl are descendants of player.
- first_name, last_name and nickname are children of player.
- first_name, last_name, nickname and teams are siblings of each other, preceding-sibling or following-sibling depending on the relative positions (XML) or alphabetical order (JSON} of their names:
- first_name is a preceding-sibling of last_name.
- last_name is a following-sibling of first_name.
To select first_names for all players with a nickname child:
To select all first_names with a nickname following-sibling (this will return the same values as
Sometimes more complex XPath manipulation is required to access and align JSON data.
kf:element_at( object, index )
Select an element at the position specified by index. This function is used when a field name is unnamed and is instead referenced by number (position).
For example, using the following data source:
Note: You can enter the formula manually or generate it automatically by choosing the Selection Option Select all <field_name> elements in all <parent> elements, adding blanks for missing <field_name> elements.
This function is used to “fill in” blanks for an element that is not included in every record. This is useful for data alignment.
For example, using the following data source:
Returns: Anders, Maurice, Wayne
Returns: Rocket, The Great One
As is, the data incorrectly aligns Anders with the nickname, Rocket, and Maurice with the nickname, The Great One.
kf:fill_elements to align this data correctly:
Returns: <blank>, Rocket, The Great One
Returns: 1974-1978, <blank>, 1978-1979
Note: You can enter the formula manually or generate it automatically by choosing the Selection Option Select all field names in <parent> element.
kf:names function returns the names of all fields contained in an element. This is useful when the field names identify categories such as demographics or countries.
For example, Facebook can return page fans data by gender and age. To create a list of these categories for a table column for a bar/line chart x-axis, use:
Which returns data shown in the following image:
which returns only value M.25-34
The following show more examples using the data source featured in the previous function (
Returns: name, wha
Returns: first_name, last_name, team
Note: Unnamed fields cannot be selected directly; see kf:element_at.
Select elements matching a set of values
In the XPath Expressions section, the first example selected elements by matching a specific value, where
/player[team/name='Winnipeg']/last_name selected the last_name for all /player elements with descendant team/name equal to Winnipeg.
- Use CONCAT to manipulate the XPath so you can add a variable.
- In the XPath,
/player[team/name='Winnipeg']/last_name, replace "Winnipeg", with the variable ( city ).
If city = Winnipeg, this formula returns
If city = Montreal, this formula returns
- Then use DATASOURCE to specify which data source (using the data source ID) to apply this reference to.
JSON and XML data sources: Videos
Watch these workshops for more information.