Method and software for obtaining answers to complex questions based on information retrieved from big data systems

A non-transitory computer-readable storage media storing one or more sequences of instructions which when executed cause one or more processors to:

• provide a simple query family that groups a plurality of first simple queries, the simple query family defined by:
• an information source for the first simple queries,
• an entity type for which features should be generated,
• first attributes that are available as filters, and
• second attributes that are usable for calculating new features;
• provide a complex query family that groups a plurality of first complex queries, the complex query family being dependent on the simple query family and the complex query family defined by:
• the information source for the first simple queries,
• the entity type for which the features should be generated, and
• third attributes that are available as filters that, at least in part, include results of one of the first simple queries;
• perform the one of the first simple queries on the information source, by:
• identifying the simple query family,
• determining the entity type defined for the simple query family,
• preprocessing data in the information source defined for the simple query family to construct a list of data retrieved from the information source according to the entity type, and
• evaluating in one batch each first simple query in the simple query family including the one of the first simple queries, using the list of data;
• obtain results of the performing of the one of the first simple queries;
• use the results to construct one of the third attributes that is available as a filter for the complex query family;
• perform one of the first complex queries using the one of the third attributes.