如何在Pandas数据框列中打印真实值的单独组的第一个和最后一个索引

Question

I wrote a program that analyzes HVAC data for operational faults. The program feeds the input data through a set of rules, and the output is a Pandas dataframe like this one.

From that output, I use this code to iterate through each column, print the name of the column itself, and print the values from the index (Date) wherever a value in the other column is true:

    pos = 0
    for column in df:
        try:
            colname = faults[df.columns[pos]]
            print "The fault -" +str (colname)+ "- occurred on:"
        except Exception:
            pass
        try:
            print df.loc[df[column] == True, 'Date'].iloc[:]
        except TypeError:
            pass
        print
        pos += 1

That output looks like this.

The code works fine, but I want to change the output a bit. I want to print just the first and last true values so that the output says something like "the fault occurred from 'x' to 'y'" instead of printing every time a true value occurs.

The complicated part is that sometimes there may be multiple groups of 1's in a column, so I can't just print the first and last indices where there are true values. A column could look like (0,0,1,1,1,0,0,0,1,1,1,1,1,1,0,0,1,0), in which case I would want it to print "the fault occurred from here to here, here to here, and here."

Is there a way to print the first and last indices of each group of true values in a Pandas dataframe column?

Answer 1

here is my suggestion, itterate through the lists to find starts and ends (add first and last if neeeded) and zip them:

df = pd.DataFrame()
df['rule_1'] = [0]*13
df['rule_2'] = [0,0,1,1,1,0,0,0,1,1,1,1,0]
df['rule_3'] = [1]*13
df.index = pd.date_range("2017-12-25 00:00",  "2017-12-25 03:00", 
freq='0.25H')

for col in df.columns:   
    starts = [i for i,x in list(enumerate(df[col].values))[1:-1] if 
((x==1)&(df[col].values[i-1]==0))]
    ends =  [i for i,x in list(enumerate(df[col].values))[1:-1] if 
((x==1)&(df[col].values[i+1]==0))]
    if df[col].values[0]==1:
        starts = [0]+starts
    if df[col].values[-1]==1:
        ends = ends + [-1]
    print (col)
    for x in zip(df.index[starts], df.index[ends]):
        print(x)
    print()

output:

rule_1

rule_2

(Timestamp('2017-12-25 00:30:00'), Timestamp('2017-12-25 01:00:00'))

(Timestamp('2017-12-25 02:00:00'), Timestamp('2017-12-25 02:45:00'))

rule_3

(Timestamp('2017-12-25 00:00:00'), Timestamp('2017-12-25 03:00:00'))