When i use pandas read_csv, pandas add some little value to the dataframe, it went from -0.079257 to -0.07925700000000001, why is this happening and how can I fix this? It also only happen to some specific values, while others seems fine.
I've tried using float_precision but seems doesn't do anything, I'm new to pandas
df = pd.read_csv('filepath')
print(df.iat[0,0])
Dataset Link
I changed the dataset file type from txt to csv manually using notepad.
Dataset Image
This is because your original data have a np.float32 precision.
import pandas as pd
df = pd.read_csv('./avila/avila-ts.txt')
print(df.iat[0,0]) # 0.13029200000000002
# stored as np.float32
df.to_csv('./my.csv',float_format=np.float32, index_label=False)
df_1 = pd.read_csv('./my.csv')
print(df_1.iat[0,0]) # 0.13029200000000002
# stored as np.float16
df.to_csv('./my.csv',float_format=np.float16, index_label=False)
df_1 = pd.read_csv('./my.csv')
print(df_1.iat[0,0]) # 0.1302
I don't know what your data is structured. could you open the data and check, better still screenshot.
data = pandas.read_csv('filepath')
data.head()
Related
I am new to Pandas in Python and I am having some difficulties returning the second column of a dataframe without column names just numbers as indexes.
import pandas as pd
import os
directory = 'A://'
sample = 'test.txt'
# Test with Air Sample
fileAir = os.path.join(directory,sample)
dataAir = pd.read_csv(fileAir,skiprows=3)
print(dataAir.iloc[:,1])
The data I am working with would be similar to:
data = [[1,2,3],[1,2,3],[1,2,3]]
Then, using pandas I wanted to have only
[[2,2,2]].
You can use
dataframe_name[column_index].values
like
df[1].values
or
dataframe_name['column_name'].values
like
df['col1'].values
I'm working on a mechanical engineering project. For the following code, the user enters the number of cylinders that their compressor has. A dataframe is then created with the correct number of columns and is exported to Excel as a CSV file.
The outputted dataframe looks exactly like I want it to as shown in the first link, but when opened in Excel it looks like the image in the second link:
1.my dataframe
2.Excel Table
Why is my dataframe not exporting properly to Excel and what can I do to get the same dataframe in Excel?
import pandas as pd
CylinderNo=int(input('Enter CylinderNo: '))
new_number=CylinderNo*3
list1=[]
for i in range(1,CylinderNo+1):
for j in range(0,3):
Cylinder_name=str('CylinderNo ')+str(i)
list1.append(Cylinder_name)
df = pd.DataFrame(list1,columns =['Kurbel/Zylinder'])
list2=['Triebwerk', 'Packung','Ventile']*CylinderNo
Bauteil = {'Bauteil': list2}
df2 = pd.DataFrame (Bauteil, columns = ['Bauteil'])
new=pd.concat([df, df2], axis=1)
list3=['Nan','Nan','Nan']*CylinderNo
Bewertung={'Bewertung': list3}
df3 = pd.DataFrame (Bewertung, columns = ['Bewertung'])
new2=pd.concat([new, df3], axis=1)
Empfehlung={'Empfehlung': list3}
df4 = pd.DataFrame (Empfehlung, columns = ['Empfehlung'])
new3=pd.concat([new2, df4], axis=1)
new3.set_index('Kurbel/Zylinder')
new3 = new3.set_index('Kurbel/Zylinder', append=True).swaplevel(0,1)
#export dataframe to csv
new3.to_csv('new3.csv')
To be clear, a comma-separated values (CSV) file is not an Excel format type or table. It is a delimited text file that Excel like other applications can open.
What you are comparing is simply presentation. Both data frames are exactly the same. For multindex data frames, Pandas print output does not repeat index values for readability on the console or IDE like Jupyter. But such values are not removed from underlying data frame only its presentation. If you re-order indexes, you will see this presentation changes. The full complete data frame is what is exported to CSV. And ideally for data integrity, you want the full data set exported with to_csv to be import-able back into Pandas with read_csv (which can set indexes) or other languages and applications.
Essentially, CSV is an industry format to store and transfer data. Consider using Excel spreadsheets, HTML markdown, or other reporting formats for your presentation needs. Therefore, to_csv may not be the best method. You can try to build text file manually with Python i/o write methods, with open('new.csv', 'w') as f, but will be an extensive workaround See also #Jeff's answer here but do note the latter part of solution does remove data.
EDIT: This was Excel's fault changing the data type, not Pandas.
When I read a CSV using pd.read_csv(file) a column of super long ints gets converted to a low res float. These ints are a date time in microseconds.
example:
CSV Columns of some values:
15555071095204000
15555071695202000
15555072295218000
15555072895216000
15555073495207000
15555074095206000
15555074695212000
15555075295202000
15555075895210000
15555076495216000
15555077095230000
15555077695206000
15555078295212000
15555078895218000
15555079495209000
15555080095208000
15555080530515000
15555086531880000
15555092531889000
15555098531886000
15555104531886000
15555110531890000
15555116531876000
15555122531873000
15555128531884000
15555134531884000
15555140531887000
15555146531874000
pd.read_csv produces: 1.55551e+16
how do I get it to report the exact int?
I've tried using: float_precision='high'
It's possible that this is caused by the way Pandas handles missing values, meaning that your column is importing as floats, to allow the missing values to be coded as NaN.
A simple solution would be to force the column to import as a str, then impute or remove missing values, and the convert to int:
import pandas as pd
df = pd.read_csv(file, dtypes={'col1': str}) # Edit to use appropriate column reference
# If you want to just remove rows with missing values, something like:
df = df[df.col1 != '']
# Then convert to integer
df.col1 = df.col1.astype('int64')
With a Minimal, Complete and Verifiable Example we can pinpoint the problem and update the code to accurately solve it.
I am reading an xlsx file using Python's Pandas pd.read_excel(myfile.xlsx,sheet_name="my_sheet",header=2) and writing the df to a csv file using df.to_csv.
The excel file contains several columns with percentage values in it (e.g. 27.44 %). In the dataframe the values are getting converted to 0.2744, i don't want any modification in data. How can i achieve this?
I already tried:
using lambda function to convert back value from 0.2744 to 27.44 % but this i don't want this because the column names/index are not fixed. It can be any col contain the % values
df = pd.read_excel(myexcel.xlsx,sheet_name="my_sheet",header=5,dtype={'column_name':str}) - Didn't work
df = pd.read_excel(myexcel.xlsx,sheet_name="my_sheet",header=5,dtype={'column_name':object}) - Didn't work
Tried xlrd module also, but that too converted % values to float.
df = pd.read_excel(myexcel.xlsx,sheet_name="my_sheet")
df.to_csv(mycsv.csv,sep=",",index=False)
from your xlsx save the file directly in csv format
To import your csv file use pandas library as follow:
import pandas as pd
df=pd.read_csv('my_sheet.csv') #in case your file located in the same directory
more information on pandas.read_csv
I have a pandas data frame 'df', it looks like below but original data has many rows.
I would like to save this as .mat file with a name 'meta.mat'. I tried;
import scipy.io as sio
sio.savemat(os.path.join(destination_folder_path,'meta.mat'), df)
This creates the meta.mat file but it only writes the field names, when I open it in matlab it looks like this;
How can I fix this, thanks.
I don't think you can pass a pd.DataFrame directly when scipy.io.savemat is expecting a dict of numpy arrays. Try replacing df with the following in your call to savemat:
{name: col.values for name, col in df.items()}
This is another solution. The resulting mat file will in the form of a structure in matlab
# data dictionary
OutData = {}
# convert DF to dictionary before loading to your dictionary
OutData['Obj'] = df.to_dict('list')
sio.savemat('path\\testmat.mat',OutData)