background

Occasionally, the content of writing code properly. It took a lot of time to get confused when specifying the dataframe in pandas. I'd like to organize it. It's compiled by pandas beginners, so if you make a mistake or have a better summary, please give me some advice.

Click here for the blog that we run: Effort 1mm

Each version is as follows.

pandas (0.18.1)
numpy (1.11.0)
Python 2.7.10

Method of specifying three elements

There are three ways to specify elements in pandas DataFrame.

df[ a ]
df.loc[ a, b ]
df.iloc[a, b]

The prerequisite DataFrame was created with the following code by referring to 10minites to pandas.

dates = pd.date_range('20130101', periods=6) df = pd.DataFrame(np.random.randn(6,4), index=dates, columns=list('ABCD'))

The resulting DataFrame looks like this.

                   A         B         C         D
2013-01-01 -0.682002  1.977886  0.348623  0.405755
2013-01-02  0.085698  2.067378 -0.356269  1.349520
2013-01-03  0.058207 -0.539280  0.023205  1.154293
2013-01-04 -0.319075  1.174168 -1.282305  0.359333
2013-01-05 -2.557677  0.922672  0.202042  0.171645
2013-01-06  1.039422  0.300340  0.701594 -0.229087

How to specify df [a] type

It seems that it can be processed in the row direction or the column direction by the specified method.

#Specify a single column
df[‘A’]  #Column name=Specify A
df.A     #Same as above

#Row direction slice: df[ 0:3 ]
df[0:3]                    #Specify from line 0 to line 3
df[‘20130102’:’20130104’]  #Index is 2013-01-02~2013-01-Specify up to 04

Specify by label name of index column-> loc method

The first of the specified arguments (may I call it?) Is the operation on the index, and the second is the operation on the column.

#Get the corresponding index
# A    0.469112
# B  -0.282863
# C   -1.509059
# D  -1.135632
# Name: 2013-01-01 00:00:00, dtype: float64
df.loc[dates[0]]

#Specify the index column at the same time.
#                    A        B
# 2013-01-01  0.469112 -0.282863
# 2013-01-02  1.212112 -0.173215
# 2013-01-03 -0.861849 -2.104569
# 2013-01-04  0.721555 -0.706771
# 2013-01-05 -0.424972  0.567020
# 2013-01-06 -0.673690  0.113648
df.loc[:, [‘A’, ‘B’]]

#Get index by index name
#                    A        B
# 2013-01-02  1.212112 -0.173215
# 2013-01-03 -0.861849 -2.104569
# 2013-01-04  0.721555 -0.706771
df.loc[‘20130102’:’20130104’, [‘A’, ‘B’]]  #When specifying multiple columns, pass them as a list

#If you want to specify only one, it is faster to use at than loc
#0.46911229990718628
df.at[dates[0],'A']

Specify by the position number of the index column

How to specify by the position of the element when replacing it with a matrix. Of course, you can select more than one.

#Specified by index position (3rd line this time)
# A    0.721555
# B  -0.706771
# C  -1.039575
# D    0.271860
# Name: 2013-01-04 00:00:00, dtype: float64
df.iloc[3]

#index/Simultaneous specification of columns(Here, 3rd to 4th rows, 0th to 1st columns)
#                   A        B
#2013-01-04  0.721555 -0.706771
#2013-01-05 -0.424972  0.567020
df.iloc[3:5,0:2]

#Specify specific elements that are skipped
#                    A        C
# 2013-01-02  1.212112  0.119209
# 2013-01-03 -0.861849 -0.494929
# 2013-01-05 -0.424972  0.276232

df.iloc[[1,2,4],[0,2]]

I have no choice but to get used to it!