Is your feature request related to a problem? Please describe.
Currently, calling len(pd.DataFrame(...)) will materialize the frame's index and compute its length.
Some storage formats (including pandas, via the PandasDataFrame object) have more efficient ways, or built-in caching mechanisms, for computing the dimensions of a frame. Adding an explicit query compiler method (get_axis_len(axis: [0, 1]) -> int) would let us take advantage of this. Accordingly, calls to len(self.index) in frontend code should be replaced with len(self), and calls to len(self.columns) with self._query_compiler.get_axis_length(1) to avoid unnecessary materialization.
Is your feature request related to a problem? Please describe. Currently, calling
len(pd.DataFrame(...))will materialize the frame's index and compute its length.Some storage formats (including pandas, via the
PandasDataFrameobject) have more efficient ways, or built-in caching mechanisms, for computing the dimensions of a frame. Adding an explicit query compiler method (get_axis_len(axis: [0, 1]) -> int) would let us take advantage of this. Accordingly, calls tolen(self.index)in frontend code should be replaced withlen(self), and calls tolen(self.columns)withself._query_compiler.get_axis_length(1)to avoid unnecessary materialization.