Another question:
How can I use ADDv2 naming conv...
# gooddata-platformp
Philippe Hébert
02/04/2022, 2:55 PMAnother question:
How can I use ADDv2 naming convention in cases where there are multiple references in a dataset 1 to a dataset 2?
Example:
dataset 1 has two fields:
r__next_appointmentr__previous_appointmentappointmentcp__appointmentb
Boris
02/04/2022, 3:06 PMI see what you are trying to do, but I am afraid that we don’t support multiple referrences between two datasets.
One way to do this would be to have separate datasets for next and previous appointments.
Other way would be have just one referrence and another attribute in the dataset 1 that would say if it’s next or previous appointment.
p
Philippe Hébert
02/04/2022, 3:08 PM😞
OK! I'll work around this somehow
Philippe Hébert
02/04/2022, 3:09 PMIt was an easy way to precompute some facts, but I'll figure out another way
Philippe Hébert
02/04/2022, 3:11 PMAnd if I want only one ref to another dataset, but I want to give it a name that's more telling of its role, can I do that?
Example:
patient.r__admission_appointmentappointment.cp__appointmentPhilippe Hébert
02/04/2022, 3:15 PMI just realized something.
If I can't have two refs from a dataset1 to a dataset2, that means I cannot have a complex calendar table that any datasetX refers to more than once, and instead I have to recreate all of the calendar logic I have as metrics?
Philippe Hébert
02/04/2022, 3:19 PMExamples of complex calendar logic:
1. Fiscal Year calculations (FY to date, FY of a date)
2. Week-of-month calculation (W1, W2, W3...) based on a most-weekday-in-month logic to determine if week is in month
b
Boris
02/04/2022, 3:47 PMHi Philippe, I am not sure how that works but anyway, in GoodData the date dimensions are predefined, you only refer to them, but it’s possible to load even custom fiscal calendars to the date dimensions - explained here: https://help.gooddata.com/doc/growth/en/dashboards-and-insights/manage-custom-date-dimensions
p
Philippe Hébert
02/04/2022, 3:51 PMGreat, thanks for that.
Philippe Hébert
02/04/2022, 3:52 PMWrt the above question:
And if I want only one ref to another dataset, but I want to give it a name that's more telling of its role, can I do that?
Example:
Can I have such a nomenclature?->patient.r__admission_appointment?appointment.cp__appointment
b
Boris
02/04/2022, 3:55 PMit’s possible to define a custom mapping in the LDM modeller, so there doesn’t have to be any relation between the name of the column in your DB and in GD, but if you want to load data based on the output stage naming convention it has to refer to the dataset name
p
Philippe Hébert
02/04/2022, 5:04 PMOK, understood. Feature is not supported in LCM ADDv2, but is supported if we decide to go deeper in the integration with custom REST API calls.
b
Boris
02/04/2022, 5:10 PMWell technically it’s supported in LCM, the custom mapping can be done only in the dev master and distributed to client projects (via release and rollout), provided it’s the same for all clients in the segment
p
Philippe Hébert
02/04/2022, 5:15 PMOh interesting. So if we configure the dev master manually, we can rollout the changes
Philippe Hébert
02/04/2022, 5:16 PMI think I'll still stick with an output-stage driven deployment and deal with the limitations this incurs
👍 1
j
Jakub Sterba
02/07/2022, 10:39 AMbtw. if you need to deal with role playing dimensions you probably need to duplicate dataset appointment with its attributes for LDM. You can perhaps use DB views for that if you do not want to duplicate data physically. You can use views also as mapping between your naming convention of columns and ADD v2 naming convention if you do not want to do the mapping in LDM.
You can refer each date to different date dataset in LDM, if you have a fact record with multiple dates. When you use a fact in Analytical Designer, the most frequently used date with the fact should be pre-selected whenever you add Date into your insight (e.g. date of next appointment). You can select the other date (e.g. previous appointment) if needed in “Date as” setting of the Date object in your insight.