Develop Reference

Splitting series of Strings into Dataframe

I have a big series of strings that I want to split into a dataframe.
The series looks like this:
s = pd.Series({"1":"Name=Marc-Age=48-Car=Tesla",
"2":"Name=Ben-Job=Pilot-Car=Porsche",
"3":"Name=Tom-Age=24-Car=Ford"})
I want to split this into a dataframe looking like this:
Name Age Job Car
1 Marc 48 Nan Tesla
2 Ben Nan Pilot Porsche
3 Tom 24 Nan Ford
I tried to split the strings first by "-" and then by "=" but don't understand how to continue after
df=s.str.split("-", expand=True)
for col in df.columns:
df[col]=df[col].str.split("=")
I get this:
` 0 1 2`
`1 ['Name', 'Marc'] ['Age', '48'] ['Car', 'Tesla']`
`2 ['Name', 'Ben'] ['Job', 'Pilot'] ['Car', 'Porsche']`
`3 ['Name', 'Tom'] ['Age', '24'] ['Car', 'Ford']`
I don't know how to continue from here. I can't loop through the rows because my dataset is really big.
Can anyone help on how to go on from here?

If you split then explode and split again you can then use a pivot.
import pandas as pd
s = pd.Series({"1":"Name=Marc-Age=48-Car=Tesla",
"2":"Name=Ben-Job=Pilot-Car=Porsche",
"3":"Name=Tom-Age=24-Car=Ford"})
s = s.str.split('-').explode().str.split('=', expand=True).reset_index()
s = s.pivot(index='index', columns=0, values=1).reset_index(drop=True)
Output
Age Car Job Name
0 48 Tesla NaN Marc
1 NaN Porsche Pilot Ben
2 24 Ford NaN Tom

Python : Remodelling a dataframe and regrouping data from a specific column with predefined rows

Let's say that I have this dataframe with four columns : "Name", "Value", "Ccy" and "Group" :
import pandas as pd
Name = ['ID', 'Country', 'IBAN','Dan_Age', 'Dan_city', 'Dan_country', 'Dan_sex', 'Dan_Age', 'Dan_country','Dan_sex' , 'Dan_city','Dan_country' ]
Value = ['TAMARA_CO', 'GERMANY','FR56','18', 'Berlin', 'GER', 'M', '22', 'FRA', 'M', 'Madrid', 'ESP']
Ccy = ['','','','EUR','EUR','USD','USD','','CHF', '','DKN','']
Group = ['0','0','0','1','1','1','1','2','2','2','3','3']
df = pd.DataFrame({'Name':Name, 'Value' : Value, 'Ccy' : Ccy,'Group':Group})
print(df)
Name Value Ccy Group
0 ID TAMARA_CO 0
1 Country GERMANY 0
2 IBAN FR56 0
3 Dan_Age 18 EUR 1
4 Dan_city Berlin EUR 1
5 Dan_country GER USD 1
6 Dan_sex M USD 1
7 Dan_Age 22 2
8 Dan_country FRA CHF 2
9 Dan_sex M 2
10 Dan_city Madrid DKN 3
11 Dan_country ESP 3
I want to represent this data differently before saving it in a csv. I would like to group the duplicates in the column "Name" with the associates values in "Values" and "Ccy". I want that the data in the column "Value" and "Ccy" are stored in the row(index) defined by the column "Group". Like that I do not mixed the data.
Then if the name is in the "group" 0, it means that it is general data so I would like that the all the rows from this "Name" are filled with the same value.
So I would like to get this result :
ID_Value Country_Value IBAN_Value Dan_age Dan_age_Ccy Dan_city_Value Dan_city_Ccy Dan_sex_Value
1 TAMARA GER FR56 18 EUR Berlin EUR M
2 TAMARA GER FR56 22 M
3 TAMARA GER FR56 Madrid DKN
I can not find how to do the first part. With the code below, I do not get what I want evn if I remove the columns empty
g = df.groupby(['Name']).cumcount()
df = df.set_index([g,'Name']).unstack().sort_index(level=1, axis=1)
df.columns = df.columns.map(lambda x: f'{x[0]}_{x[1]}')
Anyone can help me !
Thank you

You can use the following. See comments in code for each step:
s = df.loc[df['Group'] == '0', 'Name'].tolist() # this variable will be used later according to Condition 2
df['Name'] = pd.Categorical(df['Name'], categories=df['Name'].unique(), ordered=True) #this preserves order before pivoting
df = df.pivot(index='Group', columns='Name') #transforms long-to-wide per expected output
for col in df.columns:
if col[1] in s: df[col] = df[col].shift().ffill() #Condition 2
df = df.iloc[1:].replace('',np.nan).dropna(axis=1, how='all').fillna('') #dataframe cleanup
df.columns = ['_'.join(col) for col in df.columns.swaplevel()] #column name cleanup
df
Out[1]:
ID_Value Country_Value IBAN_Value Dan_Age_Value Dan_city_Value \
Group
1 TAMARA_CO GERMANY FR56 18 Berlin
2 TAMARA_CO GERMANY FR56 22
3 TAMARA_CO GERMANY FR56 Madrid
Dan_country_Value Dan_sex_Value Dan_Age_Ccy Dan_city_Ccy \
Group
1 GER M EUR EUR
2 FRA M
3 ESP DKN
Dan_country_Ccy Dan_sex_Ccy
Group
1 USD USD
2 CHF
3
From there, you can drop columns you don't want, change strings from "TAMARA_CO" to "TAMARA", "GERMANY" to "GER", use reset_index(drop=True), etc.

You can do this quite easily with only 3 steps:
Split your data frame into 2 parts: the "general data" (which we want as a series) and the more specific data. Each data frame now contains the same kinds of information.
The key part of your problem: reorganizing the data. All you need is the pandas pivot function. It does exactly what you need!
Add the general information and the pivoted data back together.
# Split Data
general = df[df.Group == "0"].set_index("Name")["Value"].copy()
main_df = df[df.Group != "0"]
# Pivot Data
result = main_df.pivot(index="Group", columns=["Name"],
values=["Value", "Ccy"]).fillna("")
result.columns = [f"{c[1]}_{c[0]}" for c in result.columns]
# Create a data frame that has an identical row for each group
general_df = pd.DataFrame([general]*3, index=result.index)
general_df.columns = [c + "_Value" for c in general_df.columns]
# Merge the data back together
result = general_df.merge(result, on="Group")
The result given above does not give the exact column order you want, so you'd have to specify that manually with
final_cols = ["ID_Value", "Country_Value", "IBAN_Value",
"Dan_age_Value", "Dan_Age_Ccy", "Dan_city_Value",
"Dan_city_Ccy", "Dan_sex_Value"]
result = result[final_cols]

Faster way to query & compute in Pandas [duplicate]

This question already has answers here:
Pandas Merging 101
(8 answers)
Closed 2 years ago.
I have two dataframes in Pandas. What I want achieve is, grab every 'Name' from DF1 and get the corresponding 'City' and 'State' present in DF2.
For example, 'Dwight' from DF1 should return corresponding values 'Miami' and 'Florida' from DF2.
DF1
Name Age Student
0 Dwight 20 Yes
1 Michael 30 No
2 Pam 55 No
. . . .
70000 Jim 27 Yes
DF1 has approx 70,000 rows with 3 columns
Second Dataframe, DF2 has approx 320,000 rows.
Name City State
0 Dwight Miami Florida
1 Michael Scranton Pennsylvania
2 Pam Austin Texas
. . . . .
325082 Jim Scranton Pennsylvania
Currently I have two functions, which return the values of 'City' and 'State' using a filter.
def read_city(id):
filt = (df2['Name'] == id)
if filt.any():
field = (df2[filt]['City'].values[0])
else:
field = ""
return field
def read_state(id):
filt = (df2['Name'] == id)
if filt.any():
field = (df2[filt]['State'].values[0])
else:
field = ""
return field
I am using the apply function to process all the values.
df['city_list'] = df['Name'].apply(read_city)
df['State_list'] = df['Name'].apply(read_state)
The result takes a long time to compute in the above way. It roughly takes me around 18 minutes to get back the df['city_list'] and df['State_list'].
Is there a faster to compute this ? Since I am completely new to pandas, I would like to know if there is a efficient way to compute this ?

I believe you can do a map:
s = df2.groupby('name')[['City','State']].agg(list)
df['city_list'] = df['Name'].map(s['City'])
df['State_list'] = df['Name'].map(s['State'])
Or a left merge after you got s:
df = df.merge(s.add_suffix('_list'), left_on='Name', right_index=True, how='left')

I think you can do something like this:
# Dataframe DF1 (dummy data)
DF1 = pd.DataFrame(columns=['Name', 'Age', 'Student'], data=[['Dwight', 20, 'Yes'], ['Michael', 30, 'No'], ['Pam', 55, 'No'], ['Jim', 27, 'Yes']])
print("DataFrame DF1")
print(DF1)
# Dataframe DF2 (dummy data)
DF2 = pd.DataFrame(columns=['Name', 'City', 'State'], data=[['Dwight', 'Miami', 'Florida'], ['Michael', 'Scranton', 'Pennsylvania'], ['Pam', 'Austin', 'Texas'], ['Jim', 'Scranton', 'Pennsylvania']])
print("DataFrame DF2")
print(DF2)
# You do a merge on 'Name' column and then, you change the name of columns 'City' and 'State'
df = pd.merge(DF1, DF2, on=['Name']).rename(columns={'City': 'city_list', 'State': 'State_list'})
print("DataFrame final")
print(df)
Output:
DataFrame DF1
Name Age Student
0 Dwight 20 Yes
1 Michael 30 No
2 Pam 55 No
3 Jim 27 Yes
DataFrame DF2
Name City State
0 Dwight Miami Florida
1 Michael Scranton Pennsylvania
2 Pam Austin Texas
3 Jim Scranton Pennsylvania
DataFrame final
Name Age Student city_list State_list
0 Dwight 20 Yes Miami Florida
1 Michael 30 No Scranton Pennsylvania
2 Pam 55 No Austin Texas
3 Jim 27 Yes Scranton Pennsylvania

Error pd.pivot "MultiIndex.name must be a hashable type"

I´m trying to apply to my pandas dataframe something similar to R's tidyr::spread . I saw in some places people using pd.pivot but so far I had no success.
So in this example I have the following dataframe DF:
df = pd.DataFrame({'action_id' : [1,2,1,4,5],
'name': ['jess', 'alex', 'jess', 'cath', 'mary'],
'address': ['house', 'house', 'park', 'park', 'park'],
'date': [ '01/01', '02/01', '03/01', '04/01', '05/01']})
How does it look like:
Ok, so what I want is a multi-index pivot table having 'action_id' and 'name' as index, "spread" the address column and fill it with the 'date' column. So my df would look like this:
What I tryed to do was:
df.pivot(index = ['action_id', 'name'], columns = 'address', values = 'date')
And I got the error TypeError: MultiIndex.name must be a hashable type
Does anyone know what am I doing wrong?

You do not need to mention the index in pd.pivot
This will work
import pandas as pd
df = pd.DataFrame({'action_id' : [1,2,1,4,5],
'name': ['jess', 'alex', 'jess', 'cath', 'mary'],
'address': ['house', 'house', 'park', 'park', 'park'],
'date': [ '01/01', '02/01', '03/01', '04/01', '05/01']})
df = pd.concat([df, pd.pivot(data=df, index=None, columns='address', values='date')], axis=1) \
.reset_index(drop=True).drop(['address','date'], axis=1)
print(df)
action_id name house park
0 1 jess 01/01 NaN
1 2 alex 02/01 NaN
2 1 jess NaN 03/01
3 4 cath NaN 04/01
4 5 mary NaN 05/01
And to arrive at what you want, you need to do a groupby
df = df.groupby(['action_id','name']).agg({'house':'first','park':'first'}).reset_index()
print(df)
action_id name house park
0 1 jess 01/01 03/01
1 2 alex 02/01 NaN
2 4 cath NaN 04/01
3 5 mary NaN 05/01
Dont forget to accept the answer if it helped you

Another option:
df2 = df.set_index(['action_id','name','address']).date.unstack().reset_index()
df2.columns.name = None

Iterate over rows and expand pandas dataframe

I have pandas dataframe with a column containing values or lists of values (of unequal length). I want to 'expand' the rows, so each value in the list becomes single value in column. An example says it all:
dfIn = pd.DataFrame({u'name': ['Tom', 'Jim', 'Claus'],
u'location': ['Amsterdam', ['Berlin','Paris'], ['Antwerp','Barcelona','Pisa'] ]})
location name
0 Amsterdam Tom
1 [Berlin, Paris] Jim
2 [Antwerp, Barcelona, Pisa] Claus
I want to turn into:
dfOut = pd.DataFrame({u'name': ['Tom', 'Jim', 'Jim', 'Claus','Claus','Claus'],
u'location': ['Amsterdam', 'Berlin','Paris', 'Antwerp','Barcelona','Pisa']})
location name
0 Amsterdam Tom
1 Berlin Jim
2 Paris Jim
3 Antwerp Claus
4 Barcelona Claus
5 Pisa Claus
I first tried using apply but it's not possible to return multiple Series as far as I know. iterrows seems to be the trick. But the code below gives me an empty dataframe...
def duplicator(series):
if type(series['location']) == list:
for location in series['location']:
subSeries = series
subSeries['location'] = location
dfOut.append(subSeries)
else:
dfOut.append(series)
for index, row in dfIn.iterrows():
duplicator(row)

Not as much interesting/fancy pandas usage, but this works:
import numpy as np
dfIn.loc[:, 'location'] = dfIn.location.apply(np.atleast_1d)
all_locations = np.hstack(dfIn.location)
all_names = np.hstack([[n]*len(l) for n, l in dfIn[['name', 'location']].values])
dfOut = pd.DataFrame({'location':all_locations, 'name':all_names})
It's about 40x faster than the apply/stack/reindex approach. As far as I can tell, that ratio holds at pretty much all dataframe sizes (didn't test how it scales with the size of the lists in each row). If you can guarantee that all location entries are already iterables, you can remove the atleast_1d call, which gives about another 20% speedup.

If you return a series whose index is a list of locations, then dfIn.apply will collate those series into a table:
import pandas as pd
dfIn = pd.DataFrame({u'name': ['Tom', 'Jim', 'Claus'],
u'location': ['Amsterdam', ['Berlin','Paris'],
['Antwerp','Barcelona','Pisa'] ]})
def expand(row):
locations = row['location'] if isinstance(row['location'], list) else [row['location']]
s = pd.Series(row['name'], index=list(set(locations)))
return s
In [156]: dfIn.apply(expand, axis=1)
Out[156]:
Amsterdam Antwerp Barcelona Berlin Paris Pisa
0 Tom NaN NaN NaN NaN NaN
1 NaN NaN NaN Jim Jim NaN
2 NaN Claus Claus NaN NaN Claus
You can then stack this DataFrame to obtain:
In [157]: dfIn.apply(expand, axis=1).stack()
Out[157]:
0 Amsterdam Tom
1 Berlin Jim
Paris Jim
2 Antwerp Claus
Barcelona Claus
Pisa Claus
dtype: object
This is a Series, while you want a DataFrame. A little massaging with reset_index gives you the desired result:
dfOut = dfIn.apply(expand, axis=1).stack()
dfOut = dfOut.to_frame().reset_index(level=1, drop=False)
dfOut.columns = ['location', 'name']
dfOut.reset_index(drop=True, inplace=True)
print(dfOut)
yields
location name
0 Amsterdam Tom
1 Berlin Jim
2 Paris Jim
3 Amsterdam Claus
4 Antwerp Claus
5 Barcelona Claus

import pandas as pd
dfIn = pd.DataFrame({
u'name': ['Tom', 'Jim', 'Claus'],
u'location': ['Amsterdam', ['Berlin','Paris'], ['Antwerp','Barcelona','Pisa'] ],
})
print(dfIn.explode('location'))
>>>
name location
0 Tom Amsterdam
1 Jim Berlin
1 Jim Paris
2 Claus Antwerp
2 Claus Barcelona
2 Claus Pisa