CODEWITHSUNDEEP
pythonpandas
JamesHudson81
JamesHudson81

Reputation: 2273

Struggle when appending rows to dataframe

I am looking forward to apply a function that involves 2 dfs:

df1:         A B C D E
   12/2/2001 3 4 2 3 4
   12/3/2001 5 5 5 4 6  
   12/4/2001 9 8 7 1 1

df_new = pd.DataFrame().reindex_like(df1)
df_new.loc[df_new.index[0]-pd.offsets.DateOffset(days=1)]=0
df_nuevo=df_new.sort_index()

for i in range(1,len(df_nuevo)):
    row=((df1.iloc[:i])*0.55)*((df_nuevo.iloc[:i-1])*0.45)
    df_nuevo.append(row)
print(df_nuevo)

The output I am expecting is df_nuevo filled with the appended rows. Actually gets filled by NaNs. Can anyone help ? Thank you .

This is the current output :

             A    B   C   D   E
   12/1/2001 0    0   0   0   0 
   12/2/2001 NaN NaN NaN NaN NaN
   12/3/2001 NaN NaN NaN NaN NaN
   12/4/2001 NaN NaN NaN NaN NaN

The idea is that where there are the NaNs appears the results of the formula specified in the part of the code : rows

Upvotes: 1

Views: 79

Answers (2)

Jonathan Eunice
Jonathan Eunice

Reputation: 22463

If you're just trying to iteratively fill out rows of df2 based on df1 and prior values of df2, a simple recasting of your formula (at least the way you originally gave it in pseudo-code) would be:

# create all-zeros df2 same shape as df1
df2 = df1.copy()
df2.loc[:,:] = 0

# iteratively compute df2
for i in range(len(df1)):
    df2.iloc[i] = df1.iloc[i] * .55 + df2.iloc[i-1] * .45

For the given example above, the results for df2 would be:

It may be instructive to see the values as they are updated through the loop:

i: 0
df1.iloc[i].values: array([ 3.,  4.,  2.,  3.,  4.])
df2.iloc[i - 1].values: array([ 0.,  0.,  0.,  0.,  0.])
resulting df2.iloc[i].values: array([ 1.65,  2.2 ,  1.1 ,  1.65,  2.2 ])

i: 1
df1.iloc[i].values: array([ 5.,  5.,  5.,  4.,  6.])
df2.iloc[i - 1].values: array([ 1.65,  2.2 ,  1.1 ,  1.65,  2.2 ])
resulting df2.iloc[i].values: array([ 3.4925,  3.74  ,  3.245 ,  2.9425,  4.29  ])

i: 2
df1.iloc[i].values: array([ 9.,  8.,  7.,  1.,  1.])
df2.iloc[i - 1].values: array([ 3.4925,  3.74  ,  3.245 ,  2.9425,  4.29  ])
resulting df2.iloc[i].values: array([ 6.521625,  6.083   ,  5.31025 ,  1.874125,  2.4805  ])

Upvotes: 1

Vaishali
Vaishali

Reputation: 38415

Is this what you are looking for? 

df1 = (df* 0.55 )+ (0.45 * df.shift())


                A       B       C       D       E
2001-12-02      NaN     NaN     NaN     NaN     NaN
2001-12-03      4.1     4.55    3.65    3.55    5.10
2001-12-04      7.2     6.65    6.10    2.35    3.25

Upvotes: 0

Related Questions