Is there a retain function in python? I want to transfer SAS code to python

Question

I want to transfer SAS code to python, and cannot find a retain function in python.

The data is like :

type_id    amount
1           100
1           200
1           400
2           0
1           200
1           300
2           0
1           150

What I want is when type_id = 2, the amount is equal to the negative value of the previous row. So the data will be like this after retain function:

type_id    amount
1           100
1           200
1           400
2          -400
1           200
1           300
2          -300
1           150

The SAS code is :

data B;
set A;
retain tempvar 0;
if type_id = 2
then amount = tempvar;
else tempvar = -amount;
drop tempvar;
run;

Does anyone have any idea about how to do this in python?
Thanks!

Asked By: juanbin

||

Source

Answer 1

IIUC

df
type_id amount
0   1   100
1   1   200
2   1   400
3   2   0
4   1   200
5   1   300
6   2   0
7   1   150

def retain(df):
    df['ret'] = df['amount'].shift()
    df.ix[df['type_id']==2,'amount'] = -df.ix[df['type_id']==2,'ret']
    df.drop("ret", axis=1, inplace=True)
    return df

retain(df)
type_id amount
0   1   100.0
1   1   200.0
2   1   400.0
3   2   -400.0
4   1   200.0
5   1   300.0
6   2   -300.0
7   1   150.0

Alternatively:

def retain(df):
    df.amount.ix[df.type_id==2] = - df.amount.shift().ix[df.type_id==2]
    return df

retain(df)
type_id amount
0   1   100.0
1   1   200.0
2   1   400.0
3   2   -400.0
4   1   200.0
5   1   300.0
6   2   -300.0
7   1   150.0

Answered By: Sergey Bushmanov

Answer 2

# another idea to try 
class retain_memo(object):
    def __init__(self):
        self.value = None
        
    def set_get(self, u):
        self.value = u 
        return u

def retain(series):
    memo = retain_memo()
    return pd.Series([memo.set_get(u) if pd.notnull(u) else memo.value for u in series])


test_series = pd.Series([1,1,2,None,None,4,4,None,None,None,None])
sas_style_retained = retain(test_series)

pd.DataFrame({'original':test_series,'retained':sas_style_retained})
# original  retained
# 0     1.0     1.0
# 1     1.0     1.0
# 2     2.0     2.0
# 3     NaN     2.0
# 4     NaN     2.0
# 5     4.0     4.0
# 6     4.0     4.0
# 7     NaN     4.0
# 8     NaN     4.0
# 9     NaN     4.0
# 10    NaN     4.0

# use something like df.groupby(<id>)[var].transform(retain) 
# for grouped data

Answered By: Jerry Watkins

Is there a retain function in python? I want to transfer SAS code to python

Question:

Answers: