Read multiple sheets in multiple excel files using pandas - python

I am trying to make a list using pandas before putting all data sets into 2D convolution layers.
And I was able to merge all data in the multiple excel files as a list.
However, the code only reads one chosen sheet name in the multiple excel files.
For example, I have 7 sheets in each excel file; named as 'gpascore1', 'gpascore2', 'gpascore3', 'gpascore4', 'gpascore5', 'gpascore6', 'gpascore7'.
And each sheet has 4 rows and 425 columns like
As shown below, you can see the code.
import os
import pandas as pd
path = os.getcwd()
files = os.listdir(path)
files_xls = [f for f in files if f[-3:] == 'xls']
df = pd.DataFrame()
for f in files_xls:
data = pd.read_excel(f, 'gpascore1') # Read only one chosen sheet available ->
gpascore1 is a sheet name.
df = df.append(data) # But there are 6 more sheets and I would like
to read data from all of the sheets
data_y = df['admit'].values
data_x = []
for i, rows in df.iterrows():
data_x.append([rows['gre'], rows['gpa'], rows['rank']])
df=df.dropna()
df.count()
Then, I got the result as below.
This is because the data from the 'gpascore1' sheet in 3 excel files were merged.
But, I want to read the data of 6 more sheets in the excel files.
Could anyone help me to find out the answer, please?
Thank you
===============<Updated code & errors>==================================
Thank you for the answers and I revised the read_excel() as
data = pd.read_excel(f, 'gpascore1') to
data = pd.read_excel(f, sheet_name=None)
But, I have key errors like below.
Could you give me any suggestions for this issue, please?
Thank you

I actually found this question under the tag of 'tensorflow'. That's hilarious. Ok, so you want to merge all Excel sheets into one dataframe?
import os
import pandas as pd
import glob
glob.glob("C:\\your_path\\*.xlsx")
all_data = pd.DataFrame()
for f in glob.glob("C:\\your_path\\*.xlsx"):
df = pd.read_excel(f)
all_data = all_data.append(df,ignore_index=True)
type(all_data)

Related

Prevent pandas from changing int to float/date?

I'm trying to merge a series of xlsx files into one, which works fine.
However, when I read a file, columns containing ints are transformed into floats (or dates?) when I merge and output them to csv. I have tried to visualize this in the picture. I have seen some solutions to this where dtype is used to "force" specific columns into int format. However, I do not always know the index nor the title of the column, so i need a more scalable solution.
Anyone with some thoughts on this?
Thank you in advance
#specify folder with xlsx-files
xlsFolder = "{}/system".format(directory)
dfMaster = pd.DataFrame()
#make a list of all xlsx-files in folder
xlsFolderContent = os.listdir(xlsFolder)
xlsFolderList = []
for file in xlsFolderContent:
if file[-5:] == ".xlsx":
xlsFolderList.append(file)
for xlsx in xlsFolderList:
print(xlsx)
xl = pd.ExcelFile("{}/{}".format(xlsFolder, xlsx))
for sheet in xl.sheet_names:
if "_Errors" in sheet:
print(sheet)
dfSheet = xl.parse(sheet)
dfSheet.fillna(0, inplace=True)
dfMaster = dfMaster.append(dfSheet)
print("len of dfMaster:", len(dfMaster))
dfMaster.to_csv("{}/dfMaster.csv".format(xlsFolder),sep=";")
Data sample:
Try to use dtype='object' as parameter of pd.read_csv or (ExcelFile.parse) to prevent Pandas to infer the data type of each column. You can also simplify your code using pathlib:
import pandas as pd
import pathlib
directory = pathlib.Path('your_path_directory')
xlsFolder = directory / 'system'
data = []
for xlsFile in xlsFolder.glob('*.xlsx'):
sheets = pd.read_excel(xlsFile, sheet_name=None, dtype='object')
for sheetname, df in sheets.items():
if '_Errors' in sheetname:
data.append(df.fillna('0'))
pd.concat(data).to_csv(xlsxFolder / dfMaster.csv, sep=';')

Python Pandas join a few files

I import a few xlsx files into pandas dataframe. It works fine, but my problem that it copies all the data under each other (so I have 10 excel file with 100 lines = 1000 lines).
I need the Dataframe with 100 lines and 10 columns, so each file will be copied next to each other and not below.
Are there any ideas how to do it?
import os
import pandas as pd
os.chdir('C:/Users/folder/')
path = ('C:/Users/folder/')
files = os.listdir(path)
allNames = pd.DataFrame()
for f in files:
info = pd.read_excel(f,'Sheet1')
allNames = allNames.append(info)
writer = pd.ExcelWriter ('Output.xlsx')
allNames.to_excel(writer, 'Copy')
writer.save()
You can feed your spreadsheets as an array of dataframes directly to pd.concat():
import os
import pandas as pd
os.chdir('C:/Users/folder/')
path = ('C:/Users/folder/')
files = os.listdir(path)
allNames = pd.concat([pd.read_excel(f,'Sheet1') for f in files], axis=1)
writer = pd.ExcelWriter ('Output.xlsx')
allNames.to_excel(writer, 'Copy')
writer.save()
Instead of stacking the tables vertically like this:
allNames = allNames.append(info)
You'll want to concatenate them horizontally like this:
allNames = pd.concat([allNames , info], axis=1)

combining excel sheets individually using pandas

I am writing a function in pandas that can read excel files from a working directory. Each of the excel files consists of multiple sheets, however the corresponding sheets in each file has the same column names and the number of sheets in each file are the same as well.
I would like to have a function that can merge/append each sheet from the different files such that sheet1 from all the files are merged into a dataframe, sheet2 from all the files are merged as second dataframe and so on. In the end, I would like to know the number of dataframes created.
For this purpose, I wrote the following code:
fpath = "/path to files/"
from os import walk
df = pd.DataFrame()
f = []
xls = []
dff = []
mypath = fpath
for (dirpath, dirnames, filenames) in walk(mypath):
f.extend(filenames)
break
for i in range(0, len(f)):
f[i] = mypath+"/"+f[i]
xls.append(pd.ExcelFile(f[i]))
cout = 0
for fil in range(0, len(xls)):
for sh in range(0, len(xls)):
if(cout <= len(xls)):
df = df.append(pd.read_excel(xls[sh], fil))
dff.append(df)
cout = cout + 1
I introduced the cout variable to control that after every merging/appending sheet 1 from all the files, the loop should break otherwise all the sheets are merged into a single dataframe.
Problem: The problem is that the function stops after returning only one dataframe in which the first sheets are merged. If I remove the "cout" switch, then all the sheets are merged. Can anyone help me in fixing the function code so that it 1)merges/append the corresponding sheets from each files, 2) make dataframe from (1), and return the dataframes? That way I will have a dataframe for each of the merged/appended sheet.
Can anyone help, Please?
Note: I am doing it in pandas but kindly suggest if you think there are better alternatives in R or any other programming language.
Ok, I looked through your code and I might have an answer for you without looping so much. Maybe it helps, maybe not.
As you point to one folder let us use listdir instead. Use pd.ExcelFile once to get the sheet names and then loop through all the sheet names and pd.concat the different excel-files for each specific sheet_name.
import pandas as pd
import os
# Preparation
p = 'exceltest' #<-- folder name
files = [os.path.join(p,i) for i in os.listdir(p) if i.endswith('.xlsx')]
sheets = pd.ExcelFile(files[0]).sheet_names
# Dictionary holding the sheet_names as keys
dfs = {s: pd.concat(pd.read_excel(f, sheet_name=s) for f in files) for s in sheets}
# Only for demo purpose
print(dfs[sheets[0]])
In my example files (named Workbook1, Workbook2) with sheet_names (Sheet 1, Sheet 2) and (Matrix A,B rowbreak 1,2) this prints:
A B
0 1 2
0 1 2

Looping through a folder to merge several excel sheets into one column

I have several workbooks, each with three sheets. I want to loop through each workbook and merge all the data from sheet_1 into a new workbook_1 file, sheet_2 into workbook_2 file & sheet_3 into workbook_3.
As far as I can tell the script below does everything I need, except rather than appending the data, it overwrites the data from the previous iteration.
For the sake of parsimony I've shortened, cleaned & simplified my script, but I'm happy to share the full script if needed.
import pandas as pd
import glob
search_dir= ('/Users/PATH/*.xlsx')
sheet_names = ['sheet_1','sheet_2','sheet_2']
def a_joiner(sheet):
for loop_x in glob.glob(search_dir):
try:
if sheet == 'sheet_1':
id_file= pd.ExcelFile(loop_x)
df_1 = id_file.parse(sheet, header= None)
writer= pd.ExcelWriter('/Users/PATH/%s.xlsx' %(sheet), engine= 'xlsxwriter')
df_1.to_excel(writer)
writer.save()
elif sheet == 'sheet_2':
#do same as above
else:
#and do same as above again
except Exception as e:
print('Error:',e)
for sheet in sheet_names:
a_joiner(sheet)
You can also easilly append data like:
df = []
for f in ['c:\\file1.xls', 'c:\\ file2.xls']:
data = pd.read_excel(f, 'Sheet1').iloc[:-2]
data.index = [os.path.basename(f)] * len(data)
df.append(data)
df = pd.concat(df)
From:
Using pandas Combining/merging 2 different Excel files/sheets

How to concatenate three excels files xlsx using python?

Hello I would like to concatenate three excels files xlsx using python.
I have tried using openpyxl, but I don't know which function could help me to append three worksheet into one.
Do you have any ideas how to do that ?
Thanks a lot
Here's a pandas-based approach. (It's using openpyxl behind the scenes.)
import pandas as pd
# filenames
excel_names = ["xlsx1.xlsx", "xlsx2.xlsx", "xlsx3.xlsx"]
# read them in
excels = [pd.ExcelFile(name) for name in excel_names]
# turn them into dataframes
frames = [x.parse(x.sheet_names[0], header=None,index_col=None) for x in excels]
# delete the first row for all frames except the first
# i.e. remove the header row -- assumes it's the first
frames[1:] = [df[1:] for df in frames[1:]]
# concatenate them..
combined = pd.concat(frames)
# write it out
combined.to_excel("c.xlsx", header=False, index=False)
I'd use xlrd and xlwt. Assuming you literally just need to append these files (rather than doing any real work on them), I'd do something like: Open up a file to write to with xlwt, and then for each of your other three files, loop over the data and add each row to the output file. To get you started:
import xlwt
import xlrd
wkbk = xlwt.Workbook()
outsheet = wkbk.add_sheet('Sheet1')
xlsfiles = [r'C:\foo.xlsx', r'C:\bar.xlsx', r'C:\baz.xlsx']
outrow_idx = 0
for f in xlsfiles:
# This is all untested; essentially just pseudocode for concept!
insheet = xlrd.open_workbook(f).sheets()[0]
for row_idx in xrange(insheet.nrows):
for col_idx in xrange(insheet.ncols):
outsheet.write(outrow_idx, col_idx,
insheet.cell_value(row_idx, col_idx))
outrow_idx += 1
wkbk.save(r'C:\combined.xls')
If your files all have a header line, you probably don't want to repeat that, so you could modify the code above to look more like this:
firstfile = True # Is this the first sheet?
for f in xlsfiles:
insheet = xlrd.open_workbook(f).sheets()[0]
for row_idx in xrange(0 if firstfile else 1, insheet.nrows):
pass # processing; etc
firstfile = False # We're done with the first sheet.
When I combine excel files (mydata1.xlsx, mydata2.xlsx, mydata3.xlsx) for data analysis, here is what I do:
import pandas as pd
import numpy as np
import glob
all_data = pd.DataFrame()
for f in glob.glob('myfolder/mydata*.xlsx'):
df = pd.read_excel(f)
all_data = all_data.append(df, ignore_index=True)
Then, when I want to save it as one file:
writer = pd.ExcelWriter('mycollected_data.xlsx', engine='xlsxwriter')
all_data.to_excel(writer, sheet_name='Sheet1')
writer.save()
Solution with openpyxl only (without a bunch of other dependencies).
This script should take care of merging together an arbitrary number of xlsx documents, whether they have one or multiple sheets. It will preserve the formatting.
There's a function to copy sheets in openpyxl, but it is only from/to the same file. There's also a function insert_rows somewhere, but by itself it won't insert any rows. So I'm afraid we are left to deal (tediously) with one cell at a time.
As much as I dislike using for loops and would rather use something compact and elegant like list comprehension, I don't see how to do that here as this is a side-effect show.
Credit to this answer on copying between workbooks.
#!/usr/bin/env python3
#USAGE
#mergeXLSX.py <a bunch of .xlsx files> ... output.xlsx
#
#where output.xlsx is the unified file
#This works FROM/TO the xlsx format. Libreoffice might help to convert from xls.
#localc --headless --convert-to xlsx somefile.xls
import sys
from copy import copy
from openpyxl import load_workbook,Workbook
def createNewWorkbook(manyWb):
for wb in manyWb:
for sheetName in wb.sheetnames:
o = theOne.create_sheet(sheetName)
safeTitle = o.title
copySheet(wb[sheetName],theOne[safeTitle])
def copySheet(sourceSheet,newSheet):
for row in sourceSheet.rows:
for cell in row:
newCell = newSheet.cell(row=cell.row, column=cell.col_idx,
value= cell.value)
if cell.has_style:
newCell.font = copy(cell.font)
newCell.border = copy(cell.border)
newCell.fill = copy(cell.fill)
newCell.number_format = copy(cell.number_format)
newCell.protection = copy(cell.protection)
newCell.alignment = copy(cell.alignment)
filesInput = sys.argv[1:]
theOneFile = filesInput.pop(-1)
myfriends = [ load_workbook(f) for f in filesInput ]
#try this if you are bored
#myfriends = [ openpyxl.load_workbook(f) for k in range(200) for f in filesInput ]
theOne = Workbook()
del theOne['Sheet'] #We want our new book to be empty. Thanks.
createNewWorkbook(myfriends)
theOne.save(theOneFile)
Tested with openpyxl 2.5.4, python 3.4.
You can simply use pandas and os library to do this.
import pandas as pd
import os
#create an empty dataframe which will have all the combined data
mergedData = pd.DataFrame()
for files in os.listdir():
#make sure you are only reading excel files
if files.endswith('.xlsx'):
data = pd.read_excel(files, index_col=None)
mergedData = mergedData.append(data)
#move the files to other folder so that it does not process multiple times
os.rename(files, 'path to some other folder')
mergedData DF will have all the combined data which you can export in a separate excel or csv file. Same code will work with csv files as well. just replace it in the IF condition
Just to add to p_barill's answer, if you have custom column widths that you need to copy, you can add the following to the bottom of copySheet:
for col in sourceSheet.column_dimensions:
newSheet.column_dimensions[col] = sourceSheet.column_dimensions[col]
I would just post this in a comment on his or her answer but my reputation isn't high enough.

Categories