Systems and methods for executing queries on tensor datasets

A method, comprising:

• maintaining, by one or more processors coupled to memory, a plurality of samples of a multi-dimensional sample dataset, each of the plurality of samples comprising a plurality of tensors, the plurality of samples comprising:
• a first sample comprising (i) a first tensor having a first value stored in a first column of the multi-dimensional sample dataset identified by a first column identifier, and (ii) a second tensor having a first plurality of values and stored in a second column of the multi-dimensional sample dataset identified by a second column identifier, and
• a second sample comprising (i) a third tensor having a second value and stored in the first column, and (ii) a fourth tensor having a second plurality of values and stored in the second column;
• identifying, by the one or more processors coupled to memory, a query for the multi-dimensional sample dataset, the query specifying aa group operation identifying the first column identifier, and (ii) a range of values to extract from tensors identified by the second column identifier;
• parsing, by the one or more processors, the query to extract the group operation and the range of values;
• executing, by the one or more processors, the query based on the group operation for the first column identifier to:
• select the first sample for a first group and the second sample for a second group based on the first value being different from the second value,
• extract, based on the range of values, a first subset of the first plurality of values of the second tensor for inclusion in the first group and a second subset of the second plurality of values of the fourth tensor for inclusion in the second group, and
• generate a set of query results comprising a first identifier of the first group and a second identifier of the second group;
• and
• providing, by the one or more processors, as output, the set of query results.