GoodData Cloud and GoodData.CN do not yet implement all capabilities that are specified on the Analytical Backend Service Provider Interface (SPI).
If you try to use any capability that is not yet available in GoodData Cloud and GoodData.CN, its implementation of the Analytical Backend will raise a
GoodData Cloud and GoodData.CN also use a different scheme for identifying entities. This impacts how you have to reference GoodData Cloud or GoodData.CN entities on input to different components or services:
- Logical data model (LDM) entities such as datasets, attributes, labels, and facts are always identified using
- Complex MAQL measures are always identified using
If you use a component or a service of the Analytical Backend that requires an instance of
ObjRef on input, you have to use the
idRef factory function to create one and specify both
NOTE: We recommend using the catalog-export tool to generate execution model
objects from your GoodData Cloud or GoodData.CN workspace. They will be generated with correct
type properties, and you can use the
model objects seamlessly as input to the different visual components.
The following capabilities are not supported on neither GoodData Cloud nor GoodData.CN:
- Combining a measure value filter and a ranking filter in a single execution
- Rollups (native totals)
- Export of execution results
- Attribute valid element queries with attribute filters or measure filters
- Dashboard alerts
- Dashboard exports to PDF and scheduling of dashboard exports
In addition to the items mentioned above, GoodData Cloud currently does not support Dashboard Plugins.
- The GoodData Cloud and GoodData.CN implementation of the Analytical Backend only supports object referencing by
type. Use the
idReffactory function to create a correct sub-type of
- The GoodData Cloud and GoodData.CN implementation of the Analytical Backend only supports attribute filtering by text values. The
newNegativeAttributeFilterfactory functions will treat inputs as text values by default.
- Drillable items used to configure drill eventing for the visual components must use
identifierto identify a drillable entity. Using
urito identify drillable items is not supported. Using
urifor drillable items may lead to undefined behavior.
uriMatchdrill predicate that can be used to configure drill eventing for visual components is not supported. You can use the
localIdentifierMatchpredicates to match measures or attributes. You can use the
attributeItemNameMatchpredicate to match an attribute element name.
- The drill events contain
uriin different parts of an event. Ignore the values in
uri. We will modify the contents of the
urifields in the upcoming releases of GoodData.UI.
Some properties can be null
Some properties related to attribute element values marked as
string can be
null if your data contains NULL values.
We will remedy that in the next major version of GoodData.UI.
In the meantime, make sure your application code can handle those fields being
null if you expect data in your workspaces to contain NULL values.
To see which fields were affected, check the commit marking them as such.