Usage¶
Xndframes is meant to provide a set of Pandas ExtensionDType/Array implementations backed by xnd
This document describes how to use the methods and classes provided by xndframes.
We will assume that the following packages have been imported.
In [1]: import xndframes as xf
In [2]: import pandas as pd
Pandas Integration¶
So far, xndframes implements XndframesArray. XndframesArray satisfies pandas extension array interface, which means that it can safely be stored inside pandas’s
Series and DataFrame.
In [3]: s = ["Pandas", "NumPy", "xnd", "SciPy", None, "CuPy", None, "Keras", "Numba"]
In [4]: packages = xf.XndframesArray(s)
In [5]: type(packages)
Out[5]: xndframes.base.XndframesArray
In [6]: print(packages.data)
xnd(['Pandas', 'NumPy', 'xnd', 'SciPy', None, 'CuPy', None, 'Keras', 'Numba'], type='9 * ?string')
In [7]: ser = pd.Series(packages)
In [8]: ser
Out[8]:
0 Pandas
1 NumPy
2 xnd
3 SciPy
4 None
5 CuPy
6 None
7 Keras
8 Numba
dtype: xndframes[9 * ?string]
In [9]: vals = list(range(9))
In [10]: values = xf.XndframesArray(vals)
In [11]: ser2 = pd.Series(values)
In [12]: ser2
Out[12]:
0 0
1 1
2 2
3 3
4 4
5 5
6 6
7 7
8 8
dtype: xndframes[9 * int64]
In [13]: df = pd.DataFrame({"packages": packages, "id": values})
In [14]: df.head()
Out[14]:
packages id
0 Pandas 0
1 NumPy 1
2 xnd 2
3 SciPy 3
4 None 4
In [15]: df
Out[15]:
packages id
0 Pandas 0
1 NumPy 1
2 xnd 2
3 SciPy 3
4 None 4
5 CuPy 5
6 None 6
7 Keras 7
8 Numba 8
In [16]: df.describe()
Out[16]:
packages id
count 7 9
unique 7 9
top NumPy 8
freq 1 1
Most pandas methods that make sense should work. The following section will call out points of interest.
In [17]: packages.shape
Out[17]: (9,)
In [18]: packages.unique()
Out[18]: <xndframes.base.XndframesArray at 0x7f0ee0262a90>
In [19]: packages.unique().data
Out[19]: xnd(['Pandas', 'NumPy', 'xnd', 'SciPy', None, 'CuPy', 'Keras', 'Numba'], type='8 * ?string')
In [20]: packages.isna()
Out[20]: array([False, False, False, False, True, False, True, False, False])
In [21]: df.info()
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 9 entries, 0 to 8
Data columns (total 2 columns):
packages 7 non-null xndframes[9 * ?string]
id 9 non-null xndframes[9 * int64]
dtypes: xndframes[9 * ?string](1), xndframes[9 * int64](1)
memory usage: 224.0 bytes
Indexing¶
If your selection returns a scalar, you get back a string.
In [22]: ser[0]
Out[22]: 'Pandas'
In [23]: df.loc[2, "packages"]
Out[23]: 'xnd'
Missing Data¶
xnd uses None to represent missing values. Xndframes does the same.
In [24]: df.isna()
Out[24]:
packages id
0 False False
1 False False
2 False False
3 False False
4 True False
5 False False
6 True False
7 False False
8 False False
In [25]: df.dropna()
Out[25]:
packages id
0 Pandas 0
1 NumPy 1
2 xnd 2
3 SciPy 3
5 CuPy 5
7 Keras 7
8 Numba 8