We Keep Coding

Pandas create column with names of columns with lowest match

I have Pandas dataframe where I have points and corresponding lengths to another points. I am able to get minimal value of the calculated columns, however, I need the column names itself. I am unable to figure out how can I get the column names corresponding to values in a new column. My dataframe looks like this:
df.head():
0 1 2 ... 6 7 min
9 58.0 94.0 984.003636 ... 696.667367 218.039561 218.039561
71 100.0 381.0 925.324708 ... 647.707783 169.856557 169.856557
61 225.0 69.0 751.353014 ... 515.152768 122.377490 122.377490
0 and 1 are datapoints, the rest are distances to datapoints #1 to 7, in some cases the number of points can differ, does not really matter for the question. The code I use to count min is following:
new = users.iloc[:,2:].min(axis=1)
users["min"] = new
#could also do the following way
#users.assign(Min=lambda users: users.iloc[:,2:].min(1))
This is quite simple and there is no much about finding the minimum of multiple columns. However, I need to get the col name instead of the value. So my desired output would look like this (in the example all are 7, which is not rule):
0 1 2 ... 6 7 min
9 58.0 94.0 984.003636 ... 696.667367 218.039561 7
71 100.0 381.0 925.324708 ... 647.707783 169.856557 7
61 225.0 69.0 751.353014 ... 515.152768 122.377490 7
Is there a simple way to achieve this?

Use df.idxmin:
In [549]: df['min'] = df.iloc[:,2:].idxmin(axis=1)
In [550]: df
Out[550]:
0 1 2 6 7 min
9 58.0 94.0 984.003636 696.667367 218.039561 7
71 100.0 381.0 925.324708 647.707783 169.856557 7
61 225.0 69.0 751.353014 515.152768 122.377490 7

Preserving id columns in dataframe after applying assign and groupby

I have a data file containing different foetal ultrasound measurements. The measurements are collected at different points during pregnancy, like so:
PregnancyID MotherID gestationalAgeInWeeks abdomCirc
0 0 14 150
0 0 21 200
1 1 20 294
1 1 25 315
1 1 30 350
2 2 8 170
2 2 9 180
2 2 18 NaN
Following this answer to a previous questions I had asked, I used this code to summarise the ultrasound measurements using the maximum measurement recorded in a single trimester (13 weeks):
(df.assign(tm = (df['gestationalAgeInWeeks']+ 13 - 1 )// 13))
.drop(columns = 'gestationalAgeInWeeks')
.groupby(['MotherID', 'PregnancyID','tm'])
.agg('max')
.unstack()
)
This results in the following output:
tm 1 2 3
MotherID PregnancyID
0 0 NaN 200.0 NaN
1 1 NaN 294.0 350.0
2 2 180.0 NaN NaN
However, MotherID and PregnancyID no longer appear as columns in the output of df.info(). Similarly, when I output the dataframe to a csv file, I only get columns 1,2 and 3. The id columns only appear when running df.head() as can be seen in the dataframe above.
I need to preserve the id columns as I want to use them to merge this dataframe with another one using the ids. Therefore, my question is, how do I preserve these id columns as part of my dataframe after running the code above?

Chain that with reset_index:
(df.assign(tm = (df['gestationalAgeInWeeks']+ 13 - 1 )// 13)
# .drop(columns = 'gestationalAgeInWeeks') # don't need this
.groupby(['MotherID', 'PregnancyID','tm'])['abdomCirc'] # change here
.max().add_prefix('abdomCirc_') # here
.unstack()
.reset_index() # and here
)
Or a more friendly version with pivot_table:
(df.assign(tm = (df['gestationalAgeInWeeks']+ 13 - 1 )// 13)
.pivot_table(index= ['MotherID', 'PregnancyID'], columns='tm',
values= 'abdomCirc', aggfunc='max')
.add_prefix('abdomCirc_') # remove this if you don't want the prefix
.reset_index()
)
Output:
tm MotherID PregnancyID abdomCirc_1 abdomCirc_2 abdomCirc_3
0 abdomCirc_0 abdomCirc_0 NaN 200.0 NaN
1 abdomCirc_1 abdomCirc_1 NaN 315.0 350.0
2 abdomCirc_2 abdomCirc_2 180.0 NaN NaN

Shifted results in pandas rolling mean

In the (5 first rows) result below, you can see Freq column and the rolling means (3) column MMeans calculated using pandas:
Freq MMeans
0 215 NaN
1 453 NaN
2 277 315.000000
3 38 256.000000
4 1 105.333333
I was expecting MMeans to start at index 1 since 1 is the mean of (0-1-2). Is there an option that I am missing with rolling method?
edit 1
print(pd.DataFrame({
'Freq':eff,
'MMeans': dF['Freq'].rolling(3).mean()}))
edit 2
Sorry #Yuca for not being as clear as I'd like to. Next is the columns I'd like pandas to return :
Freq MMeans
0 215 NaN
1 453 315.000000
2 277 256.000000
3 38 105.333333
4 1 29.666667
which are not the results returned with min_periods=2

use min_periods =1
df['rol_mean'] = df['Freq'].rolling(3,min_periods=1).mean()
output:
Freq MMeans rol_mean
0 215 NaN 215.000000
1 453 NaN 334.000000
2 277 315.000000 315.000000
3 38 256.000000 256.000000
4 1 105.333333 105.333333

Change value if consecutive number of certain condition is achieved in Pandas

I would to change the value of certain DataFrame values only if a certain condition is met an n number of consecutive times.
Example:
df = pd.DataFrame(np.random.randn(15, 3))
df.iloc[4:8,0]=40
df.iloc[12,0]=-40
df.iloc[10:12,1]=-40
Which gives me this DF:
0 1 2
0 1.238892 0.802318 -0.013856
1 -1.136326 -0.527263 -0.260975
2 1.118771 0.031517 0.527350
3 1.629482 -0.158941 -1.045209
4 40.000000 0.598657 -1.268399
5 40.000000 0.442297 -0.016363
6 40.000000 -0.316817 1.744822
7 40.000000 0.193083 0.914172
8 0.322756 -0.680682 0.888702
9 -1.204531 -0.240042 1.416020
10 -1.337494 -40.000000 -1.195780
11 -0.703669 -40.000000 0.657519
12 -40.000000 -0.288235 -0.840145
13 -1.084869 -0.298030 -1.592004
14 -0.617568 -1.046210 -0.531523
Now, if I do
a=df.copy()
a[ abs(a) > abs(a.std()) ] = float('nan')
I get
0 1 2
0 1.238892 0.802318 -0.013856
1 -1.136326 -0.527263 -0.260975
2 1.118771 0.031517 0.527350
3 1.629482 -0.158941 NaN
4 NaN 0.598657 NaN
5 NaN 0.442297 -0.016363
6 NaN -0.316817 NaN
7 NaN 0.193083 0.914172
8 0.322756 -0.680682 0.888702
9 -1.204531 -0.240042 NaN
10 -1.337494 NaN NaN
11 -0.703669 NaN 0.657519
12 NaN -0.288235 -0.840145
13 -1.084869 -0.298030 NaN
14 -0.617568 -1.046210 -0.531523
which is fair. However, I would like only to replace the values with NaN if these conditions were met by a maximum of 2 consecutive entries (so I can interpolate later). For example, I wanted the result to be
0 1 2
0 1.238892 0.802318 -0.013856
1 -1.136326 -0.527263 -0.260975
2 1.118771 0.031517 0.527350
3 1.629482 -0.158941 NaN
4 40.000000 0.598657 NaN
5 40.000000 0.442297 -0.016363
6 40.000000 -0.316817 NaN
7 40.000000 0.193083 0.914172
8 0.322756 -0.680682 0.888702
9 -1.204531 -0.240042 NaN
10 -1.337494 NaN NaN
11 -0.703669 NaN 0.657519
12 NaN -0.288235 -0.840145
13 -1.084869 -0.298030 NaN
14 -0.617568 -1.046210 -0.531523
Apparently there's no ready-to-use method to do this. The solution I found that closest resembles my problem was this one, but I couldn't make it work for me.
Any ideas?

See below - the tricky part is (cond[c] != cond[c].shift(1)).cumsum() which breaks the data into contiguous runs of the same value.
In [23]: cond = abs(df) > abs(df.std())
In [24]: for c in df.columns:
...: grouper = (cond[c] != cond[c].shift(1)).cumsum() * cond[c]
...: fill = (df.groupby(grouper)[c].transform('size') <= 2)
...: df.loc[fill, c] = np.nan
In [25]: df
Out[25]:
0 1 2
0 1.238892 0.802318 -0.013856
1 -1.136326 -0.527263 -0.260975
2 1.118771 0.031517 0.527350
3 1.629482 -0.158941 NaN
4 40.000000 0.598657 NaN
5 40.000000 0.442297 -0.016363
6 40.000000 -0.316817 NaN
7 40.000000 0.193083 0.914172
8 0.322756 -0.680682 0.888702
9 -1.204531 -0.240042 NaN
10 -1.337494 NaN NaN
11 -0.703669 NaN 0.657519
12 NaN -0.288235 -0.840145
13 -1.084869 -0.298030 NaN
14 -0.617568 -1.046210 -0.531523
To explain a bit more, cond[c] is a boolean series indicating whether your condition is true or not.
The cond[c] != cond[c].shift(1) compares the current row's condition to the next row's. This has the effecting of 'marking' where a run of values begins with the value True.
The .cumsum() converts the bools to integers and takes the cumulative sum. It may not be immediately intuitive, but this 'numbers' the groups of contiguous values. Finally the * cond[c] reassigns all groups that didn't meet the criteria to 0 (using False == 0)
So now you have groups of contiguous numbers that meet your condition, the next step performs a groupby to count how many values are in each group (transform('size').
Finally a new bool condition is used to assign missing values to those groups with 2 or less values meeting the condition.

Converting string objects to int/float using pandas

import pandas as pd
path1 = "/home/supertramp/Desktop/100&life_180_data.csv"
mydf = pd.read_csv(path1)
numcigar = {"Never":0 ,"1-5 Cigarettes/day" :1,"10-20 Cigarettes/day":4}
print mydf['Cigarettes']
mydf['CigarNum'] = mydf['Cigarettes'].apply(numcigar.get).astype(float)
print mydf['CigarNum']
mydf.to_csv('/home/supertramp/Desktop/powerRangers.csv')
The csv file "100&life_180_data.csv" contains columns like age, bmi,Cigarettes,Alocohol etc.
No int64
Age int64
BMI float64
Alcohol object
Cigarettes object
dtype: object
Cigarettes column contains "Never" "1-5 Cigarettes/day","10-20 Cigarettes/day".
I want to assign weights to these object (Never,1-5 Cigarettes/day ,....)
The expected output is new column CigarNum appended which consists only numbers 0,1,2
CigarNum is as expected till 8 rows and then shows Nan till last row in CigarNum column
0 Never
1 Never
2 1-5 Cigarettes/day
3 Never
4 Never
5 Never
6 Never
7 Never
8 Never
9 Never
10 Never
11 Never
12 10-20 Cigarettes/day
13 1-5 Cigarettes/day
14 Never
...
167 Never
168 Never
169 10-20 Cigarettes/day
170 Never
171 Never
172 Never
173 Never
174 Never
175 Never
176 Never
177 Never
178 Never
179 Never
180 Never
181 Never
Name: Cigarettes, Length: 182, dtype: object
The output I get shoudln't give NaN after few first rows.
0 0
1 0
2 1
3 0
4 0
5 0
6 0
7 0
8 0
9 0
10 NaN
11 NaN
12 NaN
13 NaN
14 0
...
167 NaN
168 NaN
169 NaN
170 NaN
171 NaN
172 NaN
173 NaN
174 NaN
175 NaN
176 NaN
177 NaN
178 NaN
179 NaN
180 NaN
181 NaN
Name: CigarNum, Length: 182, dtype: float64

OK, first problem is you have embedded spaces causing the function to incorrectly apply:
fix this using vectorised str:
mydf['Cigarettes'] = mydf['Cigarettes'].str.replace(' ', '')
now create your new column should just work:
mydf['CigarNum'] = mydf['Cigarettes'].apply(numcigar.get).astype(float)
UPDATE
Thanks to #Jeff as always for pointing out superior ways to do things:
So you can call replace instead of calling apply:
mydf['CigarNum'] = mydf['Cigarettes'].replace(numcigar)
# now convert the types
mydf['CigarNum'] = mydf['CigarNum'].convert_objects(convert_numeric=True)
you can also use factorize method also.
Thinking about it why not just set the dict values to be floats anyway and then you avoid the type conversion?
So:
numcigar = {"Never":0.0 ,"1-5 Cigarettes/day" :1.0,"10-20 Cigarettes/day":4.0}
Version 0.17.0 or newer
convert_objects is deprecated since 0.17.0, this has been replaced with to_numeric
mydf['CigarNum'] = pd.to_numeric(mydf['CigarNum'], errors='coerce')
Here errors='coerce' will return NaN where the values cannot be converted to a numeric value, without this it will raise an exception

Try using this function for all problems of this kind:
def get_series_ids(x):
'''Function returns a pandas series consisting of ids,
corresponding to objects in input pandas series x
Example:
get_series_ids(pd.Series(['a','a','b','b','c']))
returns Series([0,0,1,1,2], dtype=int)'''
values = np.unique(x)
values2nums = dict(zip(values,range(len(values))))
return x.replace(values2nums)