From xarray to xframe¶

Containers¶

Python 3 - xarray	C++ 14 - xframe
`xr.DataArray([[1, 2], [3, 4]],` `[('x', ['a', 'b']),` `('y', [1, 4])])`	`xf::variable<double>({{1, 2}, {3, 4}},` `{{"x", xf::axis({"a", "b"})},` `{"y", xf::axis({1, 4})}})`

Indexing¶

xframe returns views for multi-selection, no copy is made.

Python 3 - xarray	C++ 14 - xframe
`da[0, 1]`	`v(0, 1)`
`da.loc['a', 'Paris']`	`v.locate("a", "Paris")`
`da.isel(city=1, group=0)`	`v.iselect({{"city", 1}, {"group", 0}})`
`da.sel(city='Paris', group='a')`	`v.select({{"city", "Paris"}, {"group", "a"}})`
`da[:, 1]`	`xf::ilocate(v, xf::iall(), 1)`
`da.loc[:, 'Paris']`	`xf::locate(v, xf::all(), "Paris")`
`da.isel(city=1)`	`xf::iselect(v, {{"city", 1}})`
`da.sel(city='Paris')`	`xf::select(v, {{"city", "Paris"}})`
`da.reindex(city=['NYC', 'Paris'])`	`xf::reindex(v, {{"city", xf::axis({"NYC", "Paris"})}})`
`da.reindex_like(df)`	`xf::reindex_like(v, v2)`

Logical¶

Logical universal functions are truly lazy. xf::where(condition, a, b) does not evaluate a where condition is falsy, and it does not evaluate b where condition is truthy. xarray relies on numpy functions, that can also operate on xarray.DataArray.

Python 3 - xarray	C++ 14 - xframe
`xr.where(a > 5, a, b)` `xr.where(a > 5, 100, a)`	`xf::where(a > 5, a, b)` `xf::where(a > 5, 100, a)`
`np.any(a)`	`xf::any(a)`
`np.all(a)`	`xf::all(a)`
`np.logical_and(a, b)`	`a && b`
`np.logical_or(a, b)`	`a \|\| b`
`np.isclose(a, b)`	`xf::isclose(a, b)`
`np.allclose(a, b)`	`xf::allclose(a, b)`

Comparisons¶

xarray relies on numpy functions, that can also operate on xarray.DataArray.

Python 3 - xarray	C++ 14 - xframe
`np.equal(a, b)`	`xf::equal(a, b)`
`np.not_equal(a, b)`	`xf::not_equal(a, b)`
`np.less(a, b)`	`xf::less(a, b)` `a < b`
`np.less_equal(a, b)`	`xf::less_equal(a, b)` `a <= b`
`np.greater(a, b)`	`xf::greater(a, b)` `a > b`
`np.greater_equal(a, b)`	`xf::greater_equal(a, b)` `a >= b`

Mathematical functions¶

xframe universal functions are provided for a large set number of mathematical functions. xarray relies on numpy functions, that can also operate on xarray.DataArray.

Basic functions:

Python 3 - xarray	C++ 14 - xframe
`np.absolute(a)`	`xf::abs(a)`
`np.sign(a)`	`xf::sign(a)`
`np.remainder(a, b)`	`xf::remainder(a, b)`
`np.clip(a, min, max)`	`xf::clip(a, min, max)`
	`xf::fma(a, b, c)`

Exponential functions:

Python 3 - xarray	C++ 14 - xframe
`np.exp(a)`	`xf::exp(a)`
`np.expm1(a)`	`xf::expm1(a)`
`np.log(a)`	`xf::log(a)`
`np.log1p(a)`	`xf::log1p(a)`

Power functions:

Python 3 - xarray	C++ 14 - xframe
`np.power(a, p)`	`xf::pow(a, b)`
`np.sqrt(a)`	`xf::sqrt(a)`
`np.square(a)`	`xf::square(a)` `xf::cube(a)`
`np.cbrt(a)`	`xf::cbrt(a)`

Trigonometric functions:

Python 3 - xarray	C++ 14 - xframe
`np.sin(a)`	`xf::sin(a)`
`np.cos(a)`	`xf::cos(a)`
`np.tan(a)`	`xf::tan(a)`

Hyperbolic functions:

Python 3 - xarray	C++ 14 - xframe
`np.sinh(a)`	`xf::sinh(a)`
`np.cosh(a)`	`xf::cosh(a)`
`np.tanh(a)`	`xf::tanh(a)`

Error and gamma functions:

Python 3 - xarray	C++ 14 - xframe
`scipy.special.erf(a)`	`xf::erf(a)`
`scipy.special.gamma(a)`	`xf::tgamma(a)`
`scipy.special.gammaln(a)`	`xf::lgamma(a)`