BSR DBApi cannot share missing inputs related to identical series among different observational nodes

[Víctor de Buen Remiro]

Sometimes the same input series is used in different observational nodes.

If there are missing values in one of these input series, then DBApi will generate a different missing value for each date and each node but it is no statistical sense, due to we know that all them have exactly the same unknown value.

[Víctor de Buen Remiro]

(In [752]) All input missing variables are stored in input.missing.labels to avoid repeated declarations in BSR model definition.
Fixes #622

[Víctor de Buen Remiro]

(In [753]) All input missing variables are stored in input.missing.labels to avoid repeated declarations in BSR model definition.
Fixes #622

[Víctor de Buen Remiro]

To solve this problem you can add this parameter to your configuration set

//If it's true the automatic generated input missing variables will use 
//id_node as prefix of identifier. Else it will use MIXTURE node and all
//inputs with the same name will are referred to the same input missing
//variables. 
Real DBApi.useNodeInMissingIdentifier = False

Default value is True to ensure backward compatibility.
But you must ensure that input with the same name in different nodes have exactly the same values. Of course they could be time series between different dates but they should have the same values in the common dates.