Compute average of the pandas df conditioned on a parameter - python

I have the following df:
import numpy as np
import pandas as pd
a = []
for i in range(5):
tmp_df = pd.DataFrame(np.random.random((10,4)))
tmp_df['lvl'] = i
a.append(tmp_df)
df = pd.concat(a, axis=0)
df =
0 1 2 3 lvl
0 0.928623 0.868600 0.854186 0.129116 0
1 0.667870 0.901285 0.539412 0.883890 0
2 0.384494 0.697995 0.242959 0.725847 0
3 0.993400 0.695436 0.596957 0.142975 0
4 0.518237 0.550585 0.426362 0.766760 0
5 0.359842 0.417702 0.873988 0.217259 0
6 0.820216 0.823426 0.585223 0.553131 0
7 0.492683 0.401155 0.479228 0.506862 0
..............................................
3 0.505096 0.426465 0.356006 0.584958 3
4 0.145472 0.558932 0.636995 0.318406 3
5 0.957969 0.068841 0.612658 0.184291 3
6 0.059908 0.298270 0.334564 0.738438 3
7 0.662056 0.074136 0.244039 0.848246 3
8 0.997610 0.043430 0.774946 0.097294 3
9 0.795873 0.977817 0.780772 0.849418 3
0 0.577173 0.430014 0.133300 0.760223 4
1 0.916126 0.623035 0.240492 0.638203 4
2 0.165028 0.626054 0.225580 0.356118 4
3 0.104375 0.137684 0.084631 0.987290 4
4 0.934663 0.835608 0.764334 0.651370 4
5 0.743265 0.072671 0.911947 0.925644 4
6 0.212196 0.587033 0.230939 0.994131 4
7 0.945275 0.238572 0.696123 0.536136 4
8 0.989021 0.073608 0.720132 0.254656 4
9 0.513966 0.666534 0.270577 0.055597 4
I am learning neat pandas functionality and thus wondering, what is the easiest way to compute average along lvl column?
What I mean is:
(df[df.lvl ==0 ] + df[df.lvl ==1 ] + df[df.lvl ==2 ] + df[df.lvl ==3 ] + df[df.lvl ==4 ]) / 5
The desired output should be a table of shape (10,4), without the column lvl, where each element is the average of 5 elements (with lvl = [0,1,2,3,4]. I hope it helps.

I think need:
np.random.seed(456)
a = []
for i in range(5):
tmp_df = pd.DataFrame(np.random.random((10,4)))
tmp_df['lvl'] = i
a.append(tmp_df)
df = pd.concat(a, axis=0)
#print (df)
df1 = (df[df.lvl ==0 ] + df[df.lvl ==1 ] +
df[df.lvl ==2 ] + df[df.lvl ==3 ] +
df[df.lvl ==4 ]) / 5
print (df1)
0 1 2 3 lvl
0 0.411557 0.520560 0.578900 0.541576 2
1 0.253469 0.655714 0.532784 0.620744 2
2 0.468099 0.576198 0.400485 0.333533 2
3 0.620207 0.367649 0.531639 0.475587 2
4 0.699554 0.548005 0.683745 0.457997 2
5 0.322487 0.316137 0.489660 0.362146 2
6 0.430058 0.159712 0.631610 0.641141 2
7 0.399944 0.511944 0.346402 0.754591 2
8 0.400190 0.373925 0.340727 0.407988 2
9 0.502879 0.399614 0.321710 0.715812 2
df = df.set_index('lvl')
df2 = df.groupby(df.groupby('lvl').cumcount()).mean()
print (df2)
0 1 2 3
0 0.411557 0.520560 0.578900 0.541576
1 0.253469 0.655714 0.532784 0.620744
2 0.468099 0.576198 0.400485 0.333533
3 0.620207 0.367649 0.531639 0.475587
4 0.699554 0.548005 0.683745 0.457997
5 0.322487 0.316137 0.489660 0.362146
6 0.430058 0.159712 0.631610 0.641141
7 0.399944 0.511944 0.346402 0.754591
8 0.400190 0.373925 0.340727 0.407988
9 0.502879 0.399614 0.321710 0.715812
EDIT:
If each subset of DataFrame have index from 0 to len(subset):
df2 = df.mean(level=0)
print (df2)
0 1 2 3 lvl
0 0.411557 0.520560 0.578900 0.541576 2
1 0.253469 0.655714 0.532784 0.620744 2
2 0.468099 0.576198 0.400485 0.333533 2
3 0.620207 0.367649 0.531639 0.475587 2
4 0.699554 0.548005 0.683745 0.457997 2
5 0.322487 0.316137 0.489660 0.362146 2
6 0.430058 0.159712 0.631610 0.641141 2
7 0.399944 0.511944 0.346402 0.754591 2
8 0.400190 0.373925 0.340727 0.407988 2
9 0.502879 0.399614 0.321710 0.715812 2

The groupby function is exactly what you want. It will group based on a condition, in this case where 'lvl' is the same, and then apply the mean function to the values for each column in that group.
df.groupby('lvl').mean()

it seems like you want to group by the index and take average of all the columns except lvl
i.e.
df.groupby(df.index)[[0,1,2,3]].mean()
For a dataframe generated using
np.random.seed(456)
a = []
for i in range(5):
tmp_df = pd.DataFrame(np.random.random((10,4)))
tmp_df['lvl'] = i
a.append(tmp_df)
df = pd.concat(a, axis=0)
df.groupby(df.index)[[0,1,2,3]].mean()
outputs:
0 1 2 3
0 0.411557 0.520560 0.578900 0.541576
1 0.253469 0.655714 0.532784 0.620744
2 0.468099 0.576198 0.400485 0.333533
3 0.620207 0.367649 0.531639 0.475587
4 0.699554 0.548005 0.683745 0.457997
5 0.322487 0.316137 0.489660 0.362146
6 0.430058 0.159712 0.631610 0.641141
7 0.399944 0.511944 0.346402 0.754591
8 0.400190 0.373925 0.340727 0.407988
9 0.502879 0.399614 0.321710 0.715812
which is identical to the output from
df.groupby(df.groupby('lvl').cumcount()).mean()
without resorting to double groupby.
IMO this is cleaner to read and will for large dataframe, will be much faster.

Related

Faster way to count occurrences of values over a certain value in a column of lists in pandas?

I have this dataFrame. Each row of column data is a list containing around 50 data points. And I want to count the number of occurrences of numbers over 50 and over 20.
>>> df['data'].head(10)
0 [33.23, 51.02, 32.01 ...
1 [99.04, 38.06, 39.57...
2 [96.04, 96.72, 401.93...
3 [96.64, 99.15, 99.83...
4 [96.71, 38.93, 53.02....
5 [88.72, 37.61, 39.61...
6 [38.93, 88.72, 37.31...
7 [88.72, 39.61, 35.71...
8 [97.44, 99.04, 88.56....
9 [00.14, 89.61, 39.95...
If we transform the df to dic, it would look like below:
>>> df.to_dict()
{'data': {'row1': [33.23, 51.02, 32.01,...], 'row2': [99.04, 38.06, 39.57,...],'row3': [96.04, 96.72, 401.93,...],'row4'...}}
The expected result i would like to get is a new column called result and it stores the count of values in data column over 50.0 or over 20.0 if no values are over 50.0:
>>> df.show()
data result
0 [33.23, 51.02, 32.01 ... 1
1 [99.04, 38.06, 39.57... 1
2 [96.04, 96.72, 401.93... 3
3 [96.64, 99.15, 99.83... 3
4 [96.71, 38.93, 53.02.... 2
This is the method i used:
pandas_data_frame[result_column] = pandas_data_frame.apply(
lambda row: count_values(row[data]), axis=1)
def count__values(numlist):
count1 = sum(
x >= 50.0 for x in list)
count2 = sum(
x >= 20.0 for x in list)
return count1 if count1 > 0 else count2
However the dataframe can be extremely huge and i was wondering if there is any pandas method to improve the performance? Thanks.
Try with explode and groupby:
df[[">=50", ">=20"]] = (df.explode("data")
.groupby(level=0)["data"]
.agg([lambda x: x.ge(50).sum(),
lambda x: x.ge(20).sum()]
)
)
>>> df
data >=50 >=20
0 [33.23, 51.02, 32.01] 1 3
1 [99.04, 38.06, 39.57] 1 3
2 [96.04, 96.72, 401.93] 3 3
3 [96.64, 99.15, 99.83] 3 3
4 [96.71, 38.93, 53.02] 2 3
5 [88.72, 37.61, 39.61] 1 3
6 [38.93, 88.72, 37.31] 1 3
7 [88.72, 39.61, 35.71] 1 3
8 [97.44, 99.04, 88.56] 3 3
9 [0.14, 89.61, 39.95] 1 2

Drop rows if value in column changes

Assume I have the following pandas data frame:
my_class value
0 1 1
1 1 2
2 1 3
3 2 4
4 2 5
5 2 6
6 2 7
7 2 8
8 2 9
9 3 10
10 3 11
11 3 12
I want to identify the indices of "my_class" where the class changes and remove n rows after and before this index. The output of this example (with n=2) should look like:
my_class value
0 1 1
5 2 6
6 2 7
11 3 12
My approach:
# where class changes happen
s = df['my_class'].ne(df['my_class'].shift(-1).fillna(df['my_class']))
# mask with `bfill` and `ffill`
df[~(s.where(s).bfill(limit=1).ffill(limit=2).eq(1))]
Output:
my_class value
0 1 1
5 2 6
6 2 7
11 3 12
One of possible solutions is to:
Make use of the fact that the index contains consecutive integers.
Find index values where class changes.
For each such index generate a sequence of indices from n-2
to n+1 and concatenate them.
Retrieve rows with indices not in this list.
The code to do it is:
ind = df[df['my_class'].diff().fillna(0, downcast='infer') == 1].index
df[~df.index.isin([item for sublist in
[ range(i-2, i+2) for i in ind ] for item in sublist])]
my_class = np.array([1] * 3 + [2] * 6 + [3] * 3)
cols = np.c_[my_class, np.arange(len(my_class)) + 1]
df = pd.DataFrame(cols, columns=['my_class', 'value'])
df['diff'] = df['my_class'].diff().fillna(0)
idx2drop = []
for i in df[df['diff'] == 1].index:
idx2drop += range(i - 2, i + 2)
print(df.drop(idx_drop)[['my_class', 'value']])
Output:
my_class value
0 1 1
5 2 6
6 2 7
11 3 12

How to calculate amounts that row values greater than a specific value in pandas?

How to calculate amounts that row values greater than a specific value in pandas?
For example, I have a Pandas DataFrame dff. I want to count row values greater than 0.
dff = pd.DataFrame(np.random.randn(9,3),columns=['a','b','c'])
dff
a b c
0 -0.047753 -1.172751 0.428752
1 -0.763297 -0.539290 1.004502
2 -0.845018 1.780180 1.354705
3 -0.044451 0.271344 0.166762
4 -0.230092 -0.684156 -0.448916
5 -0.137938 1.403581 0.570804
6 -0.259851 0.589898 0.099670
7 0.642413 -0.762344 -0.167562
8 1.940560 -1.276856 0.361775
I am using an inefficient way. How to be more efficient?
dff['count'] = 0
for m in range(len(dff)):
og = 0
for i in dff.columns:
if dff[i][m] > 0:
og += 1
dff['count'][m] = og
dff
a b c count
0 -0.047753 -1.172751 0.428752 1
1 -0.763297 -0.539290 1.004502 1
2 -0.845018 1.780180 1.354705 2
3 -0.044451 0.271344 0.166762 2
4 -0.230092 -0.684156 -0.448916 0
5 -0.137938 1.403581 0.570804 2
6 -0.259851 0.589898 0.099670 2
7 0.642413 -0.762344 -0.167562 1
8 1.940560 -1.276856 0.361775 2
You can create a boolean mask of your DataFrame, that is True wherever a value is greater than your threshold (in this case 0), and then use sum along the first axis.
dff.gt(0).sum(1)
0 1
1 1
2 2
3 2
4 0
5 2
6 2
7 1
8 2
dtype: int64

Divide part of a dataframe by another while keeping columns that are not being divided

I have two data frames as below:
Sample_name C14-Cer C16-Cer C18-Cer C18:1-Cer C20-Cer
0 1 1 0.161456 0.033139 0.991840 2.111023 0.846197
1 1 10 0.636140 1.024235 36.333741 16.074662 3.142135
2 1 13 0.605840 0.034337 2.085061 2.125908 0.069698
3 1 14 0.038481 0.152382 4.608259 4.960007 0.162162
4 1 5 0.035628 0.087637 1.397457 0.768467 0.052605
5 1 6 0.114375 0.020196 0.220193 7.662065 0.077727
Sample_name C14-Cer C16-Cer C18-Cer C18:1-Cer C20-Cer
0 1 1 0.305224 0.542488 66.428382 73.615079 10.342252
1 1 10 0.814696 1.246165 73.802644 58.064363 11.179206
2 1 13 0.556437 0.517383 50.555948 51.913547 9.412299
3 1 14 0.314058 1.148754 56.165767 61.261950 9.142128
4 1 5 0.499129 0.460813 40.182454 41.770906 8.263437
5 1 6 0.300203 0.784065 47.359506 52.841821 9.833513
I want to divide the numerical values in the selected cells of the first by the second and I am using the following code:
df1_int.loc[:,'C14-Cer':].div(df2.loc[:,'C14-Cer':])
However, this way I lose the information from the column "Sample_name".
C14-Cer C16-Cer C18-Cer C18:1-Cer C20-Cer
0 0.528977 0.061088 0.014931 0.028677 0.081819
1 0.780831 0.821909 0.492309 0.276842 0.281070
2 1.088785 0.066367 0.041243 0.040951 0.007405
3 0.122529 0.132650 0.082047 0.080964 0.017738
4 0.071381 0.190178 0.034778 0.018397 0.006366
5 0.380993 0.025759 0.004649 0.145000 0.007904
How can I perform the division while keeping the column "Sample_name" in the resulting dataframe?
You can selectively overwrite using loc, the same way that you're already performing the division:
df1_int.loc[:,'C14-Cer':] = df1_int.loc[:,'C14-Cer':].div(df2.loc[:,'C14-Cer':])
This preserves the sample_name col:
In [12]:
df.loc[:,'C14-Cer':] = df.loc[:,'C14-Cer':].div(df1.loc[:,'C14-Cer':])
df
Out[12]:
Sample_name C14-Cer C16-Cer C18-Cer C18:1-Cer C20-Cer
index
0 1 1 0.528975 0.061087 0.014931 0.028677 0.081819
1 1 10 0.780831 0.821910 0.492309 0.276842 0.281070
2 1 13 1.088785 0.066367 0.041243 0.040951 0.007405
3 1 14 0.122528 0.132650 0.082047 0.080964 0.017738
4 1 5 0.071380 0.190179 0.034778 0.018397 0.006366
5 1 6 0.380992 0.025758 0.004649 0.145000 0.007904

Best way to split a DataFrame given an edge

Suppose I have the following DataFrame:
a b
0 A 1.516733
1 A 0.035646
2 A -0.942834
3 B -0.157334
4 A 2.226809
5 A 0.768516
6 B -0.015162
7 A 0.710356
8 A 0.151429
And I need to group it given the "edge B"; that means the groups will be:
a b
0 A 1.516733
1 A 0.035646
2 A -0.942834
3 B -0.157334
4 A 2.226809
5 A 0.768516
6 B -0.015162
7 A 0.710356
8 A 0.151429
That is any time I find a 'B' in the column 'a' I want to split my DataFrame.
My current solution is:
#create the dataframe
s = pd.Series(['A','A','A','B','A','A','B','A','A'])
ss = pd.Series(np.random.randn(9))
dff = pd.DataFrame({"a":s,"b":ss})
#my solution
count = 0
ls = []
for i in s:
if i=="A":
ls.append(count)
else:
ls.append(count)
count+=1
dff['grpb']=ls
and I got the dataframe:
a b grpb
0 A 1.516733 0
1 A 0.035646 0
2 A -0.942834 0
3 B -0.157334 0
4 A 2.226809 1
5 A 0.768516 1
6 B -0.015162 1
7 A 0.710356 2
8 A 0.151429 2
Which I can then split with dff.groupby('grpb').
Is there a more efficient way to do this using pandas' functions?
here's a oneliner:
zip(*dff.groupby(pd.rolling_median((1*(dff['a']=='B')).cumsum(),3,True)))[-1]
[ 1 2
0 A 1.516733
1 A 0.035646
2 A -0.942834
3 B -0.157334,
1 2
4 A 2.226809
5 A 0.768516
6 B -0.015162,
1 2
7 A 0.710356
8 A 0.151429]
How about:
df.groupby((df.a == "B").shift(1).fillna(0).cumsum())
For example:
>>> df
a b
0 A -1.957118
1 A -0.906079
2 A -0.496355
3 B 0.552072
4 A -1.903361
5 A 1.436268
6 B 0.391087
7 A -0.907679
8 A 1.672897
>>> gg = list(df.groupby((df.a == "B").shift(1).fillna(0).cumsum()))
>>> pprint.pprint(gg)
[(0,
a b
0 A -1.957118
1 A -0.906079
2 A -0.496355
3 B 0.552072),
(1, a b
4 A -1.903361
5 A 1.436268
6 B 0.391087),
(2, a b
7 A -0.907679
8 A 1.672897)]
(I didn't bother getting rid of the indices; you could use [g for k, g in df.groupby(...)] if you liked.)
An alternative is:
In [36]: dff
Out[36]:
a b
0 A 0.689785
1 A -0.374623
2 A 0.517337
3 B 1.549259
4 A 0.576892
5 A -0.833309
6 B -0.209827
7 A -0.150917
8 A -1.296696
In [37]: dff['grpb'] = np.NaN
In [38]: breaks = dff[dff.a == 'B'].index
In [39]: dff['grpb'][breaks] = range(len(breaks))
In [40]: dff.fillna(method='bfill').fillna(len(breaks))
Out[40]:
a b grpb
0 A 0.689785 0
1 A -0.374623 0
2 A 0.517337 0
3 B 1.549259 0
4 A 0.576892 1
5 A -0.833309 1
6 B -0.209827 1
7 A -0.150917 2
8 A -1.296696 2
Or using itertools to create 'grpb' is an option too.
def vGroup(dataFrame, edgeCondition, groupName='autoGroup'):
groupNum = 0
dataFrame[groupName] = ''
#loop over each row
for inx, row in dataFrame.iterrows():
if edgeCondition[inx]:
dataFrame.ix[inx, groupName] = 'edge'
groupNum += 1
else:
dataFrame.ix[inx, groupName] = groupNum
return dataFrame[groupName]
vGroup(df, df[0] == ' ')

Categories

Resources