This writer sends tables into a GoodData project and can be used with either your own GoodData account or with an account provided by Keboola. Before configuring the GoodData writer, it is important to understand that GoodData analytics relies on a Logical Data Model (LDM) of your dataset. The GoodData writer creates the LDM for you, so that you don’t have to use CloudConnect or other tools provided by GoodData. However, you need to provide all the required information in the configuration. This makes the configuration non-trivial, and you should have the data model designed before you start configuring the writer.
Create a new configuration of the Facebook Ads extractor.
The first step is to configure the GoodData project that will be used for writing:
There are several options in the configuration dialog:
Here you need to enter a project name and choose the type of the project. You can
After you set up the project, you can see the Go To Project button. This will allow you to login to the GoodData project from within your Keboola Connection project using Single Sign-On (SSO). (SSO access works only for projects provisioned by Keboola. If you used an existing project, you will have to login on your own.)
Date dimension is an important concept of the GoodData LDM. It is the question of the LDM design to determine for which columns a date dimension should be created. One rule of thumb is that different date columns in a single table should have different date dimensions. On the other hand, you might want to share a date dimension between different tables — a main date representing an event of something. It is also possible to have different date dimensions representing, e.g., a calendar year and a fiscal year.
When creating a date dimension, you have to enter the dimension name and a dimension template. The available templates are:
my-company-dimensionas Template ID. For the latter case, enter
customas the Template ID.
Click the New Table button to add a new table to the writer:
Select an existing table from the Storage and enter a Title which will be used for the table in the GoodData project (you can change the title later if needed). The next step is to configure the table columns.
Important: There is a hard limit on the GoodData interface (Workspace loading limit; link requires authentication) which allows to load only 25 GB of CSV data in a single load of a single table. If you’re nearing this limit, we suggest you split the table into chunks and load it incrementally. Feel free to contact our support for more help.
Setting the table columns involves configuration of three things:
Data Type of the column. You can choose from
DECIMAL for numbers,
VARCHAR for texts.
Consult the GoodData documentation for
the limits of each data type.
IGNORE— The column is excluded from the load and will not be created in the GoodData project.
FACT— The Fact component — a numerical piece of arbitrary data used to define metrics, e.g. Price column.
ATTRIBUTE— The Attribute component of the LDM model — a discrete set of alphanumeric or numeric data, e.g. Eye Color column containing values Blue, Brown and Green.
CONNECTION_POINT— The Connection point component — acts like a primary key, i.e. it is used to identify the rows of the table.
REFERENCE— A column which defines a Relation between tables. You have to select a Reference column as a target to which the relation points to. The target must be a
CONNECTION_POINTalready defined in another table.
DATE— A date column — you have to select a corresponding date dimension for it.
LABEL— A Label attribute which allows you to display alternative values for a column. You have to select a Reference column, which is the name of an attribute column in the same table. The label column is used as a secondary view of the reference column in the GoodData UI. For example, you can create a reference column
FullNameand add a label column
Surname. You can define multiple labels for a single attribute.
HYPERLINK— A Hyperlink attribute, which acts as a link in the reports. You have to select a Reference which is the name of an attribute column in the same table. The reference column is the one which will be used as a label, the hyperlink column is the one which is expected to contain the address.
Save your configuration.
You can also select a load type: Full Load (the default), Automatic Incremental Load or Manual Incremental Load. Incremental load will keep the existing data in the GoodData project. It can be much faster, but the source data needs to be correctly prepared. The incremental load relies on the following two features:
CONNECTION_POINTcolumn (which acts as a database primary key), or a Fact Grain (which acts as a compound unique key). Fact Grain can be used when there is no single identifying column (i.e. there is no
CONNECTION_POINT) and there is a combination of columns which can be used to identify rows. If there is no such combination, then incremental loading can’t be used.
From the table configuration page, you can also Run a load of a single table. However, if the table has relations to other
tables, it will fail with the message:
Schema reference of column 'X' of dataset Y is invalid.
There are multiple options how you can write data to a GoodData project:
By default, the LDM is updated with each run of the writer. If your project structure settled down and does not
change often, you may wish to switch to Load data only mode. This saves some time from each run and also
is somewhat safer against inadvertent changes of the model. When you change the LDM definition in the writer and load data
without updating the model, you’re likely to receive an error similar to this:
"The object identifier "XY" mentioned in the manifest
does not represent a valid object. Adjust the manifest or create the object.".
Another option is to switch between Separate load and Multiload. These represent different modes of transferring data into GoodData, and depending on the size and nature of your project, either of them may be faster.
Since the project is partially managed by the writer, there are some things you should be aware of.
The writer manages the LDM in the project. If you make changes to the LDM using other tools, they will be overwritten by the writer — unless it is used solely in Load data only mode. You can review the LDM in the GoodData project management:
The writer creates service users which are used to update the LDM and manage the SSO (
Keboola GoodData and
You should not disable those users or change their role. Any KBC project user can enter the GoodData project using
the SSO. Once they activate the SSO, a new user will be created
for them in the GoodData project (these users have
(KBC) suffix and
When you delete a writer that created a GoodData project, that project will also be deleted after a grace period. When you delete a writer that was connected to an existing project, it won’t be deleted.