other

In [2]:

# datar specific

import numpy
from datar import f
from datar.data import iris
from datar.base import as_date, factor, c
from datar.other import *
from datar.dplyr import mutate, group_by
from datar.tibble import tibble

%run nb_helpers.py
nb_header(
    # get, 
    # flatten, 
    itemgetter, 
    attrgetter, 
    pd_str, 
    pd_cat, 
    pd_dt, 
    book='datar',
)

# datar specific import numpy from datar import f from datar.data import iris from datar.base import as_date, factor, c from datar.other import * from datar.dplyr import mutate, group_by from datar.tibble import tibble %run nb_helpers.py nb_header( # get, # flatten, itemgetter, attrgetter, pd_str, pd_cat, pd_dt, book='datar', )

Try this notebook on binder.

★ itemgetter

¶

Itemgetter as a function for verb¶

In datar expression, we can do:

arr = [1,2,3]
tibble(x=2) >> mutate(y=arr[f.x])

Since arr[f.x] won't compile. We need to use the itemgetter operator:

tibble(x=2) >> mutate(y=itemgetter(arr, f.x))

Args:¶

  data: The data to be get items from
  subscr: The subscripts

★ attrgetter

¶

Attrgetter as a function for verb¶

This is helpful when we want to access to an accessor
(ie. CategoricalAccessor) from a SeriesGroupBy object

★ pd_str

¶

Pandas' str accessor for a Series (x.str)¶

This is helpful when x is a SeriesGroupBy object

★ pd_cat

¶

Pandas' cat accessor for a Series (x.cat)¶

This is helpful when x is a SeriesGroupBy object

★ pd_dt

¶

Pandas' dt accessor for a Series (x.dt)¶

This is helpful when x is a SeriesGroupBy object

In [3]:

# iris >> get(c[:5])

# iris >> get(c[:5])

In [4]:

# iris >> get(cols=f.Species)

# iris >> get(cols=f.Species)

In [5]:

# select single element
# iris >> get(1, f.Species)

# select single element # iris >> get(1, f.Species)

In [6]:

# get it as a single-element dataframe
# iris >> get([1], f.Species)

# get it as a single-element dataframe # iris >> get([1], f.Species)

In [7]:

# or 
# iris >> get(1, [f.Species])

# or # iris >> get(1, [f.Species])

In [8]:

df = tibble(x=c[1:3], y=c[3:5])
# df >> flatten()

df = tibble(x=c[1:3], y=c[3:5]) # df >> flatten()

In [12]:

arr = numpy.array(['a', 'b', 'c', 'd', 'e'])
# df >> mutate(a=arr[f.x], b=arr[f.y])  # Error
df >> mutate(a=itemgetter(arr, f.x.values), b=itemgetter(arr, f.y.values))

arr = numpy.array(['a', 'b', 'c', 'd', 'e']) # df >> mutate(a=arr[f.x], b=arr[f.y]) # Error df >> mutate(a=itemgetter(arr, f.x.values), b=itemgetter(arr, f.y.values))

Out[12]:

	x	y	a	b
	<int64>	<int64>	<object>	<object>
0	1	3	b	d
1	2	4	c	e

In [13]:

df = tibble(x=["abc", "def"])
df >> mutate(a=attrgetter(f.x, 'str').upper())

df = tibble(x=["abc", "def"]) df >> mutate(a=attrgetter(f.x, 'str').upper())

Out[13]:

	x	a
	<object>	<object>
0	abc	ABC
1	def	DEF

In [14]:

# or
# df >> mutate(a=pd_str(f.x).upper())
# or
df >> mutate(a=f.x.str.upper())

# or # df >> mutate(a=pd_str(f.x).upper()) # or df >> mutate(a=f.x.str.upper())

Out[14]:

	x	a
	<object>	<object>
0	abc	ABC
1	def	DEF

In [15]:

# but when df is grouped
gf = df >> group_by(g=[1, 2])
# pd_str(gf.x)[:2].obj
gf >> mutate(a=pd_str(gf.x)[:2])

# but when df is grouped gf = df >> group_by(g=[1, 2]) # pd_str(gf.x)[:2].obj gf >> mutate(a=pd_str(gf.x)[:2])

Out[15]:

	x	g	a
	<object>	<int64>	<object>
0	abc	1	ab
1	def	2	de

TibbleGrouped: g (n=2)

In [16]:

gf = (
    tibble(x=["2022-01-01", "2022-12-02"])
    >> mutate(x=as_date(f.x, format="%Y-%m-%d"))
    >> group_by(g=[1, 2])
)
gf >> mutate(month=pd_dt(gf.x).month)

gf = ( tibble(x=["2022-01-01", "2022-12-02"]) >> mutate(x=as_date(f.x, format="%Y-%m-%d")) >> group_by(g=[1, 2]) ) gf >> mutate(month=pd_dt(gf.x).month)

Out[16]:

	x	g	month
	<datetime64[ns]>	<int64>	<int64>
0	2022-01-01	1	1
1	2022-12-02	2	12

TibbleGrouped: g (n=2)

In [17]:

gf = (
    tibble(x=factor([1, 2], levels=[1, 2, 3]))
    >> group_by(g=[1, 2])
)
gf >> mutate(codes=pd_cat(gf.x).codes)

gf = ( tibble(x=factor([1, 2], levels=[1, 2, 3])) >> group_by(g=[1, 2]) ) gf >> mutate(codes=pd_cat(gf.x).codes)

Out[17]:

	x	g	codes
	<category>	<int64>	<int8>
0	1	1	0
1	2	2	1

TibbleGrouped: g (n=2)