I have this data frame available. It has a timestamp for start, a timestamp for end and a duration column.
start
end
duration
1
5
4
2
5
3
3
4
1
4
6
2
5
9
4
6
7
1
7
10
3
I'd like to add a column 'rolling_mean' to the dataframe that calculates a rolling average on all previous rows (ordered by start) with this condition: only previous rows can be used for mean calculation where the event has already ended (so end date should be equal to or lower than the start date of the row for which the rolling mean is being calculated). So for row number 4, the rolling_mean is 1 because we look at all previous rows and only the previous one fulfills the condition of the event having ended.
This is the dataframe I'd like to get with a Pandas rolling mean:
start
end
duration
rolling_mean
1
5
4
Nan
2
5
3
Nan
3
4
1
Nan
4
6
2
1
5
9
4
2.666667
6
7
1
2.500000
7
10
3
2.200000
Here is the code to reproduce my example:
d = [[1, 5],
[2, 5],
[3, 4],
[4, 6],
[5, 9],
[6, 7],
[7, 10]]
df = pd.DataFrame(d, columns=['start_time', 'end_time'])
df['duration'] = df.end_time - df.start_time
I've tried to merge the dataframe with itself to then filter out the irrelevant rows, but the data frame is too big to take this approach.
So I'm looking for a rolling mean but where I can specify the extra condition.
Does anyone have any ideas for this one?
A for loop will do the job:
rolling_mean = np.repeat(np.nan, len(df))
start, end, duration = df[["start_time", "end_time", "duration"]].to_numpy().T
for i in range(len(df)):
matches = duration[:i][end[:i] <= start[i]]
if matches.any():
rolling_mean[i] = matches.mean()
Related
I have following dataframe:
period symptoms recovery
1 4 2
1 5 2
1 6 2
2 3 1
2 5 2
2 8 4
2 12 6
3 4 2
3 5 2
3 6 3
3 8 5
4 5 2
4 8 4
4 12 6
I'm trying to find the common values of df['period'] groups (1, 2, 3, 4) based on value
of two columns 'symptoms' and 'recovery'
Result should be :
symptoms recovery period
5 2 [1, 2, 3, 4]
8 4 [2, 4]
where each same two columns values has the periods occurrence in a list or column.
I'm I approaching the problem in the wrong way ? Appreciate your help.
I tried to turn each period into dict and loop through to find values but didn't work for me. Also tried to use grouby().apply() but I'm not getting a meaningful data frame.
Tried sorting values based on 3 columns but couldn't get the common ones between each period section.
Last attempt :
df2 = df[['period', 'how_long', 'days_to_ex']].copy()
#s = df.groupby(["period", "symptoms", "recovery"]).size()
s = df.groupby(["symptoms", "recovery"]).size()
You were almost there:
from io import StringIO
import pandas as pd
# setup sample data
data = StringIO("""
period;symptoms;recovery
1;4;2
1;5;2
1;6;2
2;3;1
2;5;2
2;8;4
2;12;6
3;4;2
3;5;2
3;6;3
3;8;5
4;5;2
4;8;4
4;12;6
""")
df = pd.read_csv(data, sep=";")
# collect unique periods
df.groupby(['symptoms','recovery'])[['period']].agg(list).reset_index()
This gives
symptoms recovery period
0 3 1 [2]
1 4 2 [1, 3]
2 5 2 [1, 2, 3, 4]
3 6 2 [1]
4 6 3 [3]
5 8 4 [2, 4]
6 8 5 [3]
7 12 6 [2, 4]
I have got a pandas DataFrame like this:
A B
0 3 ...
1 2
2 4
3 4
4 1
5 7
6 5
7 3
I would like to compute a rolling along column A summing its elements backwards until I reach at least 10. The resulting windows should be:
A B window_indices
0 3 ... NA
1 2 NA
2 4 NA
3 4 --> [3,2,1]
4 1 [4,3,2,1]
5 7 [5,4,3]
6 5 [6,5]
7 3 [7,6,5]
Next, I want to compute some statistics on column B, something like that:
df.my_rolling(on='A', func='sum', threshold=10).B.mean()
I have got an idea: we could think of the elements of column A as seconds. Transform A in a datetime column and perform a standard rolling on it. But I don't know how to do that.
This is no able to do with rolling since the rolling window is not fixed
l = [[df.index[(df.A.loc[:x].iloc[::-1].cumsum()>=10).idxmax():x+1].tolist()[::-1]
if (df.A.loc[:x].sum()>=10) else np.nan] for x in df.A.index]
Out[46]:
[[nan],
[nan],
[nan],
[[3, 2, 1]],
[[4, 3, 2, 1]],
[[5, 4, 3]],
[[6, 5]],
[[7, 6, 5]]]
df['new'] = l
This is a multipart problem. I have found solutions for each separate part, but when I try to combine these solutions, I don't get the outcome I want.
Let's say this is my dataframe:
df = pd.DataFrame(list(zip([1, 3, 6, 7, 7, 8, 4], [6, 7, 7, 9, 5, 3, 1])), columns = ['Values', 'Vals'])
df
Values Vals
0 1 6
1 3 7
2 6 7
3 7 9
4 7 5
5 8 3
6 4 1
Let's say I want to find the pattern [6, 7, 7] in the 'Values' column.
I can use a modified version of the second solution given here:
Pandas: How to find a particular pattern in a dataframe column?
pattern = [6, 7, 7]
pat_i = [df[i-len(pattern):i] # Get the index
for i in range(len(pattern), len(df)) # for each 3 consequent elements
if all(df['Values'][i-len(pattern):i] == pattern)] # if the pattern matched
pat_i
[ Values Vals
2 6 7
3 7 9
4 7 5]
The only way I've found to narrow this down to just index values is the following:
pat_i = [df.index[i-len(pattern):i] # Get the index
for i in range(len(pattern), len(df)) # for each 3 consequent elements
if all(df['Values'][i-len(pattern):i] == pattern)] # if the pattern matched
pat_i
[RangeIndex(start=2, stop=5, step=1)]
Once I've found the pattern, what I want to do, within the original dataframe, is reorder the pattern to [7, 7, 6], moving the entire associated rows as I do this. In other words, going by the index, I want to get output that looks like this:
df.reindex([0, 1, 3, 4, 2, 5, 6])
Values Vals
0 1 6
1 3 7
3 7 9
4 7 5
2 6 7
5 8 3
6 4 1
Then, finally, I want to reset the index so that the values in all the columns stay in the new re-ordered place;
Values Vals
0 1 6
1 3 7
2 7 9
3 7 5
4 6 7
5 8 3
6 4 1
In order to use pat_i as a basis for re-ordering, I've tried to modify the second solution given here:
Python Pandas: How to move one row to the first row of a Dataframe?
target_row = 2
# Move target row to first element of list.
idx = [target_row] + [i for i in range(len(df)) if i != target_row]
However, I can't figure out how to exploit the pat_i RangeIndex object to use it with this code. The solution, when I find it, will be applied to hundreds of dataframes, each one of which will contain the [6, 7, 7] pattern that needs to be re-ordered in one place, but not the same place in each dataframe.
Any help appreciated...and I'm sure there must be an elegant, pythonic way of doing this, as it seems like it should be a common enough challenge. Thank you.
I just sort of rewrote your code. I held the first and last indexes to the side, reordered the indexes of interest, and put everything together in a new index. Then I just use the new index to reorder the data.
import pandas as pd
from pandas import RangeIndex
df = pd.DataFrame(list(zip([1, 3, 6, 7, 7, 8, 4], [6, 7, 7, 9, 5, 3, 1])), columns = ['Values', 'Vals'])
pattern = [6, 7, 7]
new_order = [1, 2, 0] # new order of pattern
for i in list(df[df['Values'] == pattern[0]].index):
if all(df['Values'][i:i+len(pattern)] == pattern):
pat_i = df[i:i+len(pattern)]
front_ind = list(range(0, pat_i.index[0]))
back_ind = list(range(pat_i.index[-1]+1, len(df)))
pat_ind = [pat_i.index[i] for i in new_order]
new_ind = front_ind + pat_ind + back_ind
df = df.loc[new_ind].reset_index(drop=True)
df
Out[82]:
Values Vals
0 1 6
1 3 7
2 7 9
3 7 5
4 6 7
5 8 3
6 4 1
Having a dataframe like this:
month transactions_ids
0 1 [0, 5, 1]
1 2 [7, 4]
2 3 [8, 10, 9, 11]
3 6 [2]
4 9 [3]
For a given transaction_id, I would like to get the month when it took place. Notice that a transaction_id can only be related to one single month.
So for example, given transaction_id = 4, the month would be 2.
I know this can be done in a loop by looking month by month if the transactions_ids related contain the given transaction_id, but I'm wondering if there is any way more efficient than that.
Cheers
The best way in my opinion is to explode your data frame and avoid having python lists in your cells.
df = df.explode('transaction_ids')
which outputs
month transactions_ids
0 1 0
0 1 5
0 1 1
1 2 7
1 2 4
2 3 8
2 3 10
2 3 9
2 3 11
3 6 2
4 9 3
Then, simply
id_to_find = 1 # example
df.loc[df.transactions_ids == id_to_find, 'month']
P.S: be aware of the duplicated indexes that explode outputs. In general, it is better to do explode(...).reset_index(drop=True) for most cases to avoid unwanted behavior.
You can use pandas string methods to find the id in the "list" (it's really just a string as far as pandas is concerned when read in using StringIO):
import pandas as pd
from io import StringIO
data = StringIO("""
month transactions_ids
1 [0,5,1]
2 [7,4]
3 [8,10,9,11]
6 [2]
9 [3]
""")
df = pd.read_csv(data, delim_whitespace=True)
df.loc[df['transactions_ids'].str.contains('4'), 'month']
In case your transactions_ids are real lists, then you can use map to check for membership:
df['transactions_ids'].map(lambda x: 3 in x)
In the spirit of Generating a list of random numbers, summing to 1 from several years ago, is there a way to apply the np array result of the np.random.dirichlet result against a groupby for the dataframe?
For example, I can loop through the unique values of the letter column and apply one at a time:
df = pd.DataFrame([['a', 1], ['a', 3], ['a', 2], ['a', 6],
['b', 7],['b', 5],['b', 4],], columns=['letter', 'value'])
df['grp_sum'] = df.groupby('letter')['value'].transform('sum')
df['prop_of_total'] = np.random.dirichlet(np.ones(len(df)), size=1).tolist()[0]
for letter in df['letter'].unique():
sz=len(df[df['letter'] == letter])
df.loc[df['letter'] == letter, 'prop_of_grp'] = np.random.dirichlet(np.ones(sz), size=1).tolist()[0]
print(df)
results in:
letter value grp_sum prop_of_total prop_of_grp
0 a 1 12 0.015493 0.293481
1 a 3 12 0.114027 0.043973
2 a 2 12 0.309150 0.160818
3 a 6 12 0.033999 0.501729
4 b 7 16 0.365276 0.617484
5 b 5 16 0.144502 0.318075
6 b 4 16 0.017552 0.064442
but there's got to be a better way than iterating the unique values and filtering the dataframe for each. This is small but I'll have potentially tens of thousands of groupings of varying sizes of ~50-100 rows each, and each needs a different random distribution.
I have also considered creating a temporary dataframe for each grouping, appending to a second dataframe and finally merging the results, though that seems more convoluted than this. I have not found a solution where I can apply an array of groupby size to the groupby but I think something along those lines would do.
Thoughts? Suggestions? Solutions?
IIUC, do a transform():
def direchlet(x, size=1):
return np.array(np.random.dirichlet(np.ones(len(x)), size=size)[0])
df['prop_of_grp'] = df.groupby('letter')['value'].transform(direchlet)
Output:
letter value grp_sum prop_of_total prop_of_grp
0 a 1 12 0.102780 0.127119
1 a 3 12 0.079201 0.219648
2 a 2 12 0.341158 0.020776
3 a 6 12 0.096956 0.632456
4 b 7 16 0.193970 0.269094
5 b 5 16 0.012905 0.516035
6 b 4 16 0.173031 0.214871