CODEWITHSUNDEEP
pythonpandasdataframe
Behdad
Behdad

Reputation: 1629

Many If statements in Pandas

I have a stock dataframe. And have many if statement to run to assign a value.

c1 = np.where((blackCondition & whitePCondition) & (close_gt_oP & open_ge_cP),'UP',None)
c2 = np.where((blackCondition & whitePCondition) & (open_lt_cP & close_le_oP),'DOWN',None)

c3 = np.where((blackCondition & whitePCondition) & (close_gt_oP & open_lt_cP),'OUTSIDE',None)
c4 = np.where((blackCondition & whitePCondition) & (close_le_oP & open_ge_cP),'INSIDE',None)
#IF2
c5 = np.where((blackCondition & blackPCondition) & (open_gt_oP & close_ge_cP),'UP',None)
c6 = np.where((blackCondition & blackPCondition) & (close_lt_cP & open_le_oP),'DOWN',None)

c7 = np.where((blackCondition & blackPCondition) & (open_gt_oP & close_lt_cP),'OUTSIDE',None)
c8 = np.where((blackCondition & blackPCondition) & (open_le_oP & close_ge_cP),'INSIDE',None)
#IF3

c9 = np.where((whiteCondition & whitePCondition) & (close_gt_cP & open_ge_oP),'UP',None)
c10 = np.where((whiteCondition & whitePCondition) & (open_lt_oP & close_le_cP),'DOWN',None)

c11 = np.where((whiteCondition & whitePCondition) & (close_gt_cP & open_lt_oP),'OUTSIDE',None)
c12 = np.where((whiteCondition & whitePCondition) & (close_le_cP & open_ge_oP),'INSIDE',None)
#IF4
c13 = np.where((whiteCondition & blackPCondition) & (open_gt_cP & close_ge_oP),'UP',None)
c14 = np.where((whiteCondition & blackPCondition) & (close_lt_oP & open_le_cP),'DOWN',None)

c15 = np.where((whiteCondition & blackPCondition) & (open_gt_cP & close_lt_oP),'OUTSIDE',None)
c16 = np.where((whiteCondition & blackPCondition) & (open_le_cP & close_ge_oP),'INSIDE',None)

How can I assign TRUE VALUE of a statement to a column?

df['pos'] = IF ANY OF THIS STATEMENT IS TRUE

ex:

if(c1) => df['pos'] = TRUE_CONDITION_OF_C1

Upvotes: 1

Views: 63

Answers (2)

jezrael
jezrael

Reputation: 862406

I think need numpy.select for set values by conditions, also I try reassign same conditions to variables m1-m3 for better performance:

m1 = (blackCondition & whitePCondition)
m2 = (blackCondition & blackPCondition)
m3 = (whiteCondition & whitePCondition)
#same like m1
#m4 = (whitePCondition & blackCondition)

c1 = m1 & (close_gt_oP & open_ge_cP)
c2 = m1 & (open_lt_cP & close_le_oP)
c5 = m2 & (open_gt_oP & close_ge_cP)
c6 = m2 & (close_lt_cP & open_le_oP)
c9 = m3 & (close_gt_cP & open_ge_oP)
c10 = m3 & (open_lt_oP & close_le_cP)
c13 = m1 & (open_gt_cP & close_ge_oP)
c14 = m1 & (close_lt_oP & open_le_cP)

c3 = m1 & (close_gt_oP & open_lt_cP)
c4 = m1 & (close_le_oP & open_ge_cP)
c7 = m2 & (open_gt_oP & close_lt_cP)
c8 = m2 & (open_le_oP & close_ge_cP)
c11 = m3 & (close_gt_cP & open_lt_oP)
c12 = m3 & (close_le_cP & open_ge_oP)
c15 = m1 & (open_gt_cP & close_lt_oP)
c16 = m1 & (open_le_cP & close_ge_oP)

Chain conditions by | for OR:

cup = c1 | c5 | c9 | c13
cdown = c2 | c6 | c10 | c14

coutside = c3 | c7 | c11 | c15
cinside = c4 | c8 | c12 | c16

df['pos'] = np.select([cup, cdown, coutside, cinside], 
                      ['UP','DOWN','OUTSIDE','INSIDE'], default=None)

Upvotes: 1

Petergavinkin
Petergavinkin

Reputation: 150

With pandas I don't know of a very elegant way of doing this. There are several options though.

You can repetitively use .loc[] with a mask, which is ugly, or you could use a .map() or .apply() method.

The following example assumes that the mask is nowhere null and that is has the same shape as the df itself.

df['pos'] = (
    (blackCondition & whitePCondition)
    & (close_gt_oP & open_ge_cP)
).map({True: "UP", False: "DOWN"})

That would only do the first line.

Upvotes: 1

Related Questions