TypeError: unsupported operand type(s) for -: 'str' and 'str' in python 3.x Anaconda

Question

I am trying to count some instances per hour time in a large dataset. The code below seems to work fine on python 2.7 but I had to upgrade it to 3.x latest version of python with all updated packages on Anaconda. When I am trying to execute the program I am getting following str error

Code:

import pandas as pd
from datetime import datetime,time
import numpy as np

fn = r'00_input.csv'
cols = ['UserId', 'UserMAC', 'HotspotID', 'StartTime', 'StopTime']
df = pd.read_csv(fn, header=None, names=cols)

df['m'] = df.StopTime + df.StartTime
df['d'] = df.StopTime - df.StartTime

# 'start' and 'end' for the reporting DF: `r`
# which will contain equal intervals (1 hour in this case)
start = pd.to_datetime(df.StartTime.min(), unit='s').date()
end = pd.to_datetime(df.StopTime.max(), unit='s').date() + pd.Timedelta(days=1)

# building reporting DF: `r`
freq = '1H'  # 1 Hour frequency
idx = pd.date_range(start, end, freq=freq)
r = pd.DataFrame(index=idx)
r['start'] = (r.index - pd.datetime(1970,1,1)).total_seconds().astype(np.int64)

# 1 hour in seconds, minus one second (so that we will not count it twice)
interval = 60*60 - 1

r['LogCount'] = 0
r['UniqueIDCount'] = 0

for i, row in r.iterrows():
        # intervals overlap test
        # https://en.wikipedia.org/wiki/Interval_tree#Overlap_test
        # i've slightly simplified the calculations of m and d
        # by getting rid of division by 2,
        # because it can be done eliminating common terms
    u = df[np.abs(df.m - 2*row.start - interval) < df.d + interval].UserID
    r.ix[i, ['LogCount', 'UniqueIDCount']] = [len(u), u.nunique()]

r['Date'] = pd.to_datetime(r.start, unit='s').dt.date
r['Day'] = pd.to_datetime(r.start, unit='s').dt.weekday_name.str[:3]
r['StartTime'] = pd.to_datetime(r.start, unit='s').dt.time
r['EndTime'] = pd.to_datetime(r.start + interval + 1, unit='s').dt.time

#r.to_csv('results.csv', index=False)
#print(r[r.LogCount > 0])
#print (r['StartTime'], r['EndTime'], r['Day'], r['LogCount'], r['UniqueIDCount'])

rout =  r[['Date', 'StartTime', 'EndTime', 'Day', 'LogCount', 'UniqueIDCount'] ]
#print rout
rout.to_csv('o_1_hour.csv', index=False, header=False

)

Where do I make changes to get a error free execution

Error:

File "C:\Program Files\Anaconda3\lib\site-packages\pandas\core\ops.py", line 686, in <lambda>
    lambda x: op(x, rvalues))

TypeError: unsupported operand type(s) for -: 'str' and 'str'

Appreciate the Help, Thanks in advance

Solution 1

I think you need change header=0 for select first row to header - then column names are replace by list cols.

If still problem, need to_numeric, because some values in StartTime and StopTime are strings, which are parsed to NaN, replace by 0 an last convert column to int:

cols = ['UserId', 'UserMAC', 'HotspotID', 'StartTime', 'StopTime']
df = pd.read_csv('canada_mini_unixtime.csv', header=0, names=cols)
#print (df)

df['StartTime'] = pd.to_numeric(df['StartTime'], errors='coerce').fillna(0).astype(int)
df['StopTime'] =  pd.to_numeric(df['StopTime'], errors='coerce').fillna(0).astype(int)

No change:

df['m'] = df.StopTime + df.StartTime
df['d'] = df.StopTime - df.StartTime
start = pd.to_datetime(df.StartTime.min(), unit='s').date()
end = pd.to_datetime(df.StopTime.max(), unit='s').date() + pd.Timedelta(days=1)

freq = '1H'  # 1 Hour frequency
idx = pd.date_range(start, end, freq=freq)
r = pd.DataFrame(index=idx)
r['start'] = (r.index - pd.datetime(1970,1,1)).total_seconds().astype(np.int64)

# 1 hour in seconds, minus one second (so that we will not count it twice)
interval = 60*60 - 1

r['LogCount'] = 0
r['UniqueIDCount'] = 0

ix is deprecated in last version of pandas, so use loc and column name is in []:

for i, row in r.iterrows():
        # intervals overlap test
        # https://en.wikipedia.org/wiki/Interval_tree#Overlap_test
        # i've slightly simplified the calculations of m and d
        # by getting rid of division by 2,
        # because it can be done eliminating common terms
    u = df.loc[np.abs(df.m - 2*row.start - interval) < df.d + interval, 'UserId']
    r.loc[i, ['LogCount', 'UniqueIDCount']] = [len(u), u.nunique()]

r['Date'] = pd.to_datetime(r.start, unit='s').dt.date
r['Day'] = pd.to_datetime(r.start, unit='s').dt.weekday_name.str[:3]
r['StartTime'] = pd.to_datetime(r.start, unit='s').dt.time
r['EndTime'] = pd.to_datetime(r.start + interval + 1, unit='s').dt.time

print (r)

Solution 2

df['d'] = df.StopTime - df.StartTime is attempting to subtract a string from another string. I don't know what your data looks like, but chances are that you want to parse StopTime and StartTime as dates. Try

df = pd.read_csv(fn, header=None, names=cols, parse_dates=[3,4])

instead of df = pd.read_csv(fn, header=None, names=cols).

Comments

[Sitz Blogz]

Thank you for the answer: When I changed the statement I am getting this error TypeError: Can't convert 'int' object to str implicitly

[Sitz Blogz]

Sure ..Thank you .. I'll make the changes and try to execute the program and let you know. Thanks again

[Sitz Blogz]

I have a bunch of code to migrate from 2.7 to 3.x .. I hope this works for me .. One down from count of pending heap also counts.. Thanks a bunch ..

[jezrael]

Yes, I starts with python 2 also, I know what is it. Good luck!

[Sitz Blogz]

Gives me error in another place ` in <module> u = df[np.abs(df.StartTime - 2*row.start - interval) < df.StopTime + interval].UserID` TypeError: incompatible type for a datetime/timedelta operation [__sub__]

[jezrael]

Hmmm, I was wrong. Ther is another problem. I edit answer.

[Sitz Blogz]

I made all the changed you mention in the answer .. But unformtunately the outputs are 0 This is a program to find log of connections every hour and unique connections in one hour.

[jezrael]

Are you sure data are OK? Is possible test it in python 2 if same output?

[Sitz Blogz]

Yes i am sure the Python 2 version works as I have generated outputs for datasets. I will work on python 2 code for the same dataset and let you know.

[Sitz Blogz]

The python 2.7 execution output in the dropbox, But rolling back and working with this dataset and the same programming isnt working .. dropbox.com/s/fcx683fgqctsqte/o_1_hour.csv?dl=0

[jezrael]

Thank you, but now I am offline only on phone. Is possible send me your python 2 code with some sample data on my email? Because still dont understand difference, code should work same way in python 2 and python 3. If data confidental, is possinle anonymize it?

[Sitz Blogz]

Sure I can share the code and input samples from a folder on dropbox, this data is already anonymous. Please can you share your email id ..

[Sitz Blogz]

Shared the dropbox repository

[jezrael]

Thank you for sharing, I think you can see new file 0_time_split_jez.py. But I am a bit confused - How is possible compare outputs if not same inputs? Also I think there is problem in something else, because no python 2 to python 3 conversion like map, print ... What do you think?

[Sitz Blogz]

The 'dart' input n output both are given, input dart is mini version and the Canada mini input is wat I want to work on now.. It might not be only python 2 or 3 but also the supporting packages like pandas, numpy and others, I did try to downgrade them aswell and check. But looks like getting the combination of same working versions a bit time taking process.. Hence i wanted to upgrade the code

[jezrael]

Ok, I try Canada mini input, but how I can check if output is correct if impossible compare with o_1_hour.csv ? Minimal there is difference in length and in dates (2010 vs 2004 years)

[Sitz Blogz]

Can I add u on Google hangout chat? Please

[jezrael]

invitation sent

[Sitz Blogz]

I did from my Gmail u have to accept it "phani.lav()gmail()com"

[jezrael]

Ok, no problem.