I am trying to read different worksheets from an Excel workbook in Python with Pandas. When I read the entire workbook and then I want to apply a .merge() then the first worksheet is read but the others are not considered. I tried to read each worksheet of the workbook but I guess they were not successfully converted to data frames because when I apply .merge() I end up with the following error: ValueError: Invalid file path or buffer object type: <class 'pandas.core.frame.DataFrame'>
This is what I have done so far:
This code works for converting the entire workbook to a data frame but only the data of the first worksheet is processed
import pandas as pd
import pypyodbc
#sql extractor
start_date = date.today()
retrieve_values = "[DEV].[CS].[QT_KPIExport] #start_date='{start_date:%Y-%m-%d}'".format(
start_date=start_date)
connection = pypyodbc.connect(driver="{SQL Server}", server="xxx.xxx.xxx.xxx", uid="X",pwd="xxx", Trusted_Connection="No")
data_frame_sql = pd.read_sql(retrieve_values, connection)
#Read the entire workbook
wb_data = pd.ExcelFile("C:\\Users\\Dev\\Testing\\Daily_Data\\NSN-Daily Data Report.xlsx")
#Convert to a dataframe the entire workbook
data_frame_excel = pd.read_excel(wb_data,index_col=None,na_values=['NA'],parse_cols="J")
#apply merge
merged_df = data_frame_sql.merge(data_frame_excel,how="inner",on="sectorname")
This code tries to read the different worksheets and convert them to data frames with no success...yet! (check the answer below)
data_frame_sql = pd.read_sql(retrieve_values, connection)
#Method 1: Tried to parse worksheet 2
#Read the entire workbook and select the specific worksheet
wb_data = pd.ExcelFile("C:\\Users\\Dev\\Testing\\Daily_Data\\NSN-Daily Data Report.xlsx", sheetname="Sheet-2")
data_frame_excel = pd.read_excel(wb_data,index_col=None,na_values=['NA'],parse_cols="J")
#apply merge
merged_df = data_frame_sql.merge(data_frame_excel,how="inner",on="sectorname")
#No success... the data of the first sheet is read
#Method 2: Tried to parse worksheet 2
#Read the entire workbook
wb_data = pd.ExcelFile("C:\\Users\\Dev\\Testing\\Daily_Data\\NSN-Daily Data Report.xlsx")
data_frame_excel = pd.read_excel(wb_data,index_col=None,na_values=['NA'],parse_cols="J")
#select one specific sheet
ws_sheet_2 = wb_data.parse("Sheet-2")
#apply merge
merged_df = data_frame_sql.merge(ws_sheet_2,how="inner",on="sectorname")
# No success.... ValueError: Invalid file path or buffer object type: <class 'pandas.core.frame.DataFrame'>
Any help or advice is greatly appreciated.
You can get all worksheets from a workbook into a dictionary by using the sheetname=None argument with the read_excel method. Key/value pairs will be ws name/dataframe.
ws_dict = pd.read_excel('excel_file.xlsx', sheetname=None)
Note the sheetname argument will change to sheet_name in future pandas versions...
I found out a solution that did the trick.
#Method 1: Add the sheetname once you have read the entire workbook
#Read the entire workbook
wb_data = pd.ExcelFile("C:\\Users\\Dev\\Testing\\Daily_Data\\NSN-Daily Data
Report.xlsx")
#Select your sheetname to read
data_frame_excel = pd.read_excel(wb_data,index_col=None,na_values=
['NA'],parse_cols="J" sheetname="Sheet-2")
#apply merge
merged_df =
data_frame_sql.merge(data_frame_excel,how="inner",on="sectorname")
To read .xlsx files in Pandas, for a document with multiple sheets, specify the sheet name and use a different engine.
Step 1 (install the openpyxl package):
! pip install openpyxl
Step 2 (use the openpyxl engine):
data_df = pd.read_excel(<ARCHIVE_PATH>, sheetname= <sheet_name>, engine='openpyxl')
Here is the official documentation.
Another solution using openpyxl directly:
wb = load_workbook(ARCHIVE_PATH)
ws = wb[<sheet-name>]
data_df = pd.DataFrame(ws.values)
df_tm = sheet.values
coluna_tm = next(df_tm)[0:]
df = pd.DataFrame(df_tm, columns=coluna_tm)
Related
I am trying to take a workbook, loop through specific worksheets retrieve a dataframe, manipulate it and essentially paste the dataframe back in the same place without changing any of the other data / sheets in the document, this is what I am trying:
path= '<folder location>.xlsx'
wb = pd.ExcelFile(path)
for sht in ['sheet1','sheet2','sheet3']:
df= pd.read_excel(wb,sheet_name = sht, skiprows = 607,nrows = 11, usecols = range(2,15))
# here I manipulate the df, to then save it down in the same place
df.to_excel(wb,sheet_name = sht, startcol=3, startrow=607)
# Save down file
wb.save(path))
wb.close()
My solution so far will just save the first sheet down with ONLY the data that I manipulated, I lose all other sheets and data that was on the sheet that I want to stay, so I end up with just sheet1 with only the data I manipulated.
Would really appreciate any help, thank you
Try using an ExcelWriter instead of an ExcelFile:
path= 'folder location.xlsx'
with pd.ExcelWriter(path) as writer:
for sht in ['sheet1','sheet2','sheet3']:
df= pd.read_excel(wb,sheet_name = sht, skiprows = 607,nrows = 11, usecols = range(2,15))
####here I manipulate the df, to then save it down in the same place###
df.to_excel(writer,sheet_name = sht, startcol=3, startrow=607)
Although I am not sure how it will behave when the file already exists and you overwrite some of them. It might be easier to read everything in first, manipulate the required sheets and save to a new file.
I am trying to merge multiple .xls files that have many columns, but 1 column with hyperlinks. I try to do this with Python but keep running into unsolvable errors.
Just to be concise, the hyperlinks are hidden under a text section. The following ctrl-click hyperlink is an example of what I encounter in the .xls files: ES2866911 (T3).
In order to improve reproducibility, I have added .xls1 and .xls2 samples below.
xls1:
Title
Publication_Number
P_A
ES2866911 (T3)
P_B
EP3887362 (A1)
.xls2:
Title
Publication_Number
P_C
AR118706 (A2)
P_D
ES2867600 (T3)
Desired outcome:
Title
Publication_Number
P_A
ES2866911 (T3)
P_B
EP3887362 (A1)
P_C
AR118706 (A2)
P_D
ES2867600 (T3)
I am unable to get .xls file into Python without losing formatting or losing hyperlinks. In addition I am unable to convert .xls files to .xlsx. I have no possibility to acquire the .xls files in .xlsx format. Below I briefly summarize what I have tried:
1.) Reading with pandas was my first attempt. Easy to do, but all hyperlinks are lost in PD, furthermore all formatting from original file is lost.
2.) Reading .xls files with openpyxl.load
InvalidFileException: openpyxl does not support the old .xls file format, please use xlrd to read this file, or convert it to the more recent .xlsx file format.
3.) Converting .xls files to .xlsx
from xls2xlsx import XLS2XLSX
x2x = XLS2XLSX(input.file.xls)
wb = x2x.to_xlsx()
x2x.to_xlsx('output_file.xlsx')
TypeError: got invalid input value of type <class 'xml.etree.ElementTree.Element'>, expected string or Element
import pyexcel as p
p.save_book_as(file_name=input_file.xls, dest_file_name=export_file.xlsx)
TypeError: got invalid input value of type <class 'xml.etree.ElementTree.Element'>, expected string or Element
During handling of the above exception, another exception occurred:
StopIteration
4.) Even if we are able to read the .xls file with xlrd for example (meaning we will never be able to save the file as .xlsx, I can't even see the hyperlink:
import xlrd
wb = xlrd.open_workbook(file) # where vis.xls is your test file
ws = wb.sheet_by_name('Sheet1')
ws.cell(5, 1).value
'AR118706 (A2)' #Which is the name, not hyperlink
5.) I tried installing older versions of openpyxl==3.0.1 to overcome type error to no succes. I tried to open .xls file with openpyxl with xlrd engine, similar typerror "xml.entree.elementtree.element' error occured. I tried many ways to batch convert .xls files to .xlsx all with similar errors.
Obviously I can just open with excel and save as .xlsx but this defeats the entire purpose, and I can't do that for 100's of files.
You need to use xlrd library to read the hyperlinks properly, pandas to merge all data together and xlsxwriter to write the data properly.
Assuming all input files have same format, you can use below code.
# imports
import os
import xlrd
import xlsxwriter
import pandas as pd
# required functions
def load_excel_to_df(filepath, hyperlink_col):
book = xlrd.open_workbook(file_path)
sheet = book.sheet_by_index(0)
hyperlink_map = sheet.hyperlink_map
data = pd.read_excel(filepath)
hyperlink_col_index = list(data.columns).index(hyperlink_col)
required_links = [v.url_or_path for k, v in hyperlink_map.items() if k[1] == hyperlink_col_index]
data['hyperlinks'] = required_links
return data
# main code
# set required variables
input_data_dir = 'path/to/input/data/'
hyperlink_col = 'Publication_Number'
output_data_dir = 'path/to/output/data/'
output_filename = 'combined_data.xlsx'
# read and combine data
required_files = os.listdir(input_data_dir)
combined_data = pd.DataFrame()
for file in required_files:
curr_data = load_excel_to_df(data_dir + os.sep + file, hyperlink_col)
combined_data = combined_data.append(curr_data, sort=False, ignore_index=True)
cols = list(combined_data.columns)
m, n = combined_data.shape
hyperlink_col_index = cols.index(hyperlink_col)
# writing data
writer = pd.ExcelWriter(output_data_dir + os.sep + output_filename, engine='xlsxwriter')
combined_data[cols[:-1]].to_excel(writer, index=False, startrow=1, header=False) # last column contains hyperlinks
workbook = writer.book
worksheet = writer.sheets[list(workbook.sheetnames.keys())[0]]
for i, col in enumerate(cols[:-1]):
worksheet.write(0, i, col)
for i in range(m):
worksheet.write_url(i+1, hyperlink_col_index, combined_data.loc[i, cols[-1]], string=combined_data.loc[i, hyperlink_col])
writer.save()
References:
reading hyperlinks - https://stackoverflow.com/a/7057076/17256762
pandas to_excel header formatting - Remove default formatting in header when converting pandas DataFrame to excel sheet
writing hyperlinks with xlsxwriter - https://xlsxwriter.readthedocs.io/example_hyperlink.html
Without a clear reproducible example, the problem is not clear. Assume I have two files called tmp.xls and tmp2.xls containing dummy data as in the two screenshots below.
Then pandas can easily, load, concatenate, and convert to .xlsx format without loss of hyperlinks. Here is some demo code and the resulting file:
import pandas as pd
f1 = pd.read_excel('tmp.xls')
f2 = pd.read_excel('tmp2.xls')
f3 = pd.concat([f1, f2], ignore_index=True)
f3.to_excel('./f3.xlsx')
Inspired by #Kunal, I managed to write code that avoids using Pandas libraries. .xls files are read by xlrd, and written to a new excel file by xlwt. Hyperlinks are maintened, and output file was saved as .xlsx format:
import os
import xlwt
from xlrd import open_workbook
# read and combine data
directory = "random_directory"
required_files = os.listdir(directory)
#Define new file and sheet to get files into
new_file = xlwt.Workbook(encoding='utf-8', style_compression = 0)
new_sheet = new_file.add_sheet('Sheet1', cell_overwrite_ok = True)
#Initialize header row, can be done with any file
old_file = open_workbook(directory+"/"+required_files[0], formatting_info=True)
old_sheet = old_file.sheet_by_index(0)
for column in list(range(0, old_sheet.ncols)):
new_sheet.write(0, column, old_sheet.cell(0, column).value) #To create header row
#Add rows from all files present in folder
for file in required_files:
old_file = open_workbook(directory+"/"+file, formatting_info=True)
old_sheet = old_file.sheet_by_index(0) #Define old sheet
hyperlink_map = old_sheet.hyperlink_map #Create map of all hyperlinks
for row in range(1, old_sheet.nrows): #We need all rows except header row
if row-1 < len(hyperlink_map.items()): #Statement to ensure we do not go out of range on the lower side of hyperlink_map.items()
Row_depth=len(new_sheet._Worksheet__rows) #We need row depth to know where to add new row
for col in list(range(old_sheet.ncols)): #For every column we need to add row cell
if col is 1: #We need to make an exception for column 2 being the hyperlinked column
click=list(hyperlink_map.items())[row-1][1].url_or_path #define URL
new_sheet.write(Row_depth, col, xlwt.Formula('HYPERLINK("{}", "{}")'.format(click, old_sheet.cell(row, 1).value)))
else: #If not hyperlinked column
new_sheet.write(Row_depth, col, old_sheet.cell(row, col).value) #Write cell
new_file.save("random_directory/output_file.xlsx")
I assume the same as daedalus in terms of the excel files. Instead of pandas I use openpyxl to read and create a new excel file.
import openpyxl
wb1 = openpyxl.load_workbook('tmp.xlsx')
ws1 = wb.get_sheet_by_name('Sheet1')
wb2 = openpyxl.load_workbook('tmp2.xlsx')
ws2 = wb.get_sheet_by_name('Sheet1')
csvDict = {}
# Go through first sheet to find the hyperlinks and keys.
for (row in ws1.max_row):
hyperlink_dict[ws1.cell(row=row, column=1).value] =
[ws1.cell(row=row, column=2).hyperlink.target,
ws1.cell(row=row, column=2).value]
# Go Through second sheet to find hyperlinks and keys.
for (row in ws2.max_row):
hyperlink_dict[ws2.cell(row=row, column=1).value] =
[ws2.cell(row=row, column=2).hyperlink.target,
ws2.cell(row=row, column=2).value]
Now you have all the data so you can create a new workbook and save the values from the dict into it via opnenpyxl.
wb = Workbook(write_only=true)
ws = wb.create_sheet()
for irow in len(csvDict):
#use ws.append() to add the data from the csv.
wb.save('new_big_file.xlsx')
https://openpyxl.readthedocs.io/en/stable/optimized.html#write-only-mode
I'm new to Python (and programming in general) and am running into a problem when writing data out to sheets in Excel.
I'm reading in an Excel file, performing a sum calculation on specific columns, and then writing the results out to a new workbook. Then at the end, it creates two charts based on the results.
The code works, except every time I run it, it creates new sheets with numbers appended to the end. I really just want it to overwrite the sheet names I provide, instead of creating new ones.
I'm not familiar enough with all the modules to understand all the options that are available. I've researched openpyxl, and pandas, and similar examples to what I'm trying to do either aren't easy to find, or don't seem to work when I try them.
import pandas as pd
import xlrd
import openpyxl as op
from openpyxl import load_workbook
import matplotlib.pyplot as plt
# declare the input file
input_file = 'TestData.xlsx'
# declare the output_file name to be written to
output_file = 'TestData_Output.xlsx'
book = load_workbook(output_file)
writer = pd.ExcelWriter(output_file, engine='openpyxl')
writer.book = book
# read the source Excel file and calculate sums
excel_file = pd.read_excel(input_file)
num_events_main = excel_file.groupby(['Column1']).sum()
num_events_type = excel_file.groupby(['Column2']).sum()
# create dataframes and write names and sums out to new workbook/sheets
df_1 = pd.DataFrame(num_events_main)
df_2 = pd.DataFrame(num_events_type)
df_1.to_excel(writer, sheet_name = 'TestSheet1')
df_2.to_excel(writer, sheet_name = 'TestSheet2')
# save and close
writer.save()
writer.close()
# dataframe for the first sheet
df = pd.read_excel(output_file, sheet_name='TestSheet1')
values = df[['Column1', 'Column3']]
# dataframe for the second sheet
df = pd.read_excel(output_file, sheet_name='TestSheet2')
values_2 = df[['Column2', 'Column3']]
# create the graphs
events_graph = values.plot.bar(x = 'Column1', y = 'Column3', rot = 60) # rot = rotation
type_graph = values_2.plot.bar(x = 'Column2', y = 'Column3', rot = 60) # rot = rotation
plt.show()
I get the expected results, and the charts work fine. I'd really just like to get the sheets to overwrite with each run.
From the pd.DataFrame.to_excel documentation:
Multiple sheets may be written to by specifying unique sheet_name.
With all data written to the file it is necessary to save the changes.
Note that creating an ExcelWriter object with a file name that already
exists will result in the contents of the existing file being erased.
Try writing to the book like
import pandas as pd
df = pd.DataFrame({'col1':[1,2,3],'col2':[4,5,6]})
writer = pd.ExcelWriter('g.xlsx')
df.to_excel(writer, sheet_name = 'first_df')
df.to_excel(writer, sheet_name = 'second_df')
writer.save()
If you inspect the workbook, you will have two worksheets.
Then lets say you wanted to write new data to the same workbook:
writer = pd.ExcelWriter('g.xlsx')
df.to_excel(writer, sheet_name = 'new_df')
writer.save()
If you inspect the workbook now, you will just have one worksheet named new_df
If there are other worksheets in the excel file that you want to keep and just overwrite the desired worksheets, you would need to use load_workbook.
Before you wrtie any data, you could delete the sheets you want to write to with:
std=book.get_sheet_by_name(<sheee_name>)
book.remove_sheet(std)
That will stop the behavior where a number gets appended to the worksheet name once you attempt to write a workbook with a duplicate sheet name.
I am trying to add an empty excel sheet into an existing Excel File using python xlsxwriter.
Setting the formula up as follows works well.
workbook = xlsxwriter.Workbook(file_name)
worksheet_cover = workbook.add_worksheet("Cover")
Output4 = workbook
Output4.close()
But once I try to add further sheets with dataframes into the Excel it overwrites the previous excel:
with pd.ExcelWriter('Luther_April_Output4.xlsx') as writer:
data_DifferingRates.to_excel(writer, sheet_name='Differing Rates')
data_DifferingMonthorYear.to_excel(writer, sheet_name='Differing Month or Year')
data_DoubleEntries.to_excel(writer, sheet_name='Double Entries')
How should I write the code, so that I can add empty sheets and existing data frames into an existing excel file.
Alternatively it would be helpful to answer how to switch engines, once I have produced the Excel file...
Thanks for any help!
If you're not forced use xlsxwriter try using openpyxl. Simply pass 'openpyxl' as the Engine for the pandas built-in ExcelWriter class. I had asked a question a while back on why this works. It is helpful code. It works well with the syntax of pd.to_excel() and it won't overwrite your already existing sheets.
from openpyxl import load_workbook
import pandas as pd
book = load_workbook(file_name)
writer = pd.ExcelWriter(file_name, engine='openpyxl')
writer.book = book
data_DifferingRates.to_excel(writer, sheet_name='Differing Rates')
data_DifferingMonthorYear.to_excel(writer, sheet_name='Differing Month or Year')
data_DoubleEntries.to_excel(writer, sheet_name='Double Entries')
writer.save()
You could use pandas.ExcelWriter with optional mode='a' argument for appending to existing Excel workbook.
You can also append to an existing Excel file:
>>> with ExcelWriter('path_to_file.xlsx', mode='a') as writer:`
... df.to_excel(writer, sheet_name='Sheet3')`
However unfortunately, this requires using a different engine, since as you observe the ExcelWriter does not support the optional mode='a' (append). If you try to pass this parameter to the constructor, it raises an error.
So you will need to use a different engine to do the append, like openpyxl. You'll need to ensure that the package is installed, otherwise you'll get a "Module Not Found" error. I have tested using openpyxl as the engine, and it is able to append new a worksheet to existing workbook:
with pd.ExcelWriter(engine='openpyxl', path='Luther_April_Output4.xlsx', mode='a') as writer:
data_DifferingRates.to_excel(writer, sheet_name='Differing Rates')
data_DifferingMonthorYear.to_excel(writer, sheet_name='Differing Month or Year')
data_DoubleEntries.to_excel(writer, sheet_name='Double Entries')
I think you need to write the data into a new file. This works for me:
# Write multiple tabs (sheets) into to a new file
import pandas as pd
from openpyxl import load_workbook
Work_PATH = r'C:\PythonTest'+'\\'
ar_source = Work_PATH + 'Test.xlsx'
Output_Wkbk = Work_PATH + 'New_Wkbk.xlsx'
# Need workbook from openpyxl load_workbook to enumerage tabs
# is there another way with only xlsxwriter?
workbook = load_workbook(filename=ar_source)
# Set sheet names in workbook as a series.
# You can also set the series manually tabs = ['sheet1', 'sheet2']
tabs = workbook.sheetnames
print ('\nWorkbook sheets: ',tabs,'\n')
# Replace this function with functions for what you need to do
def default_col_width (df, sheetname, writer):
# Note, this seems to use xlsxwriter as the default engine.
for column in df:
# map col width to col name. Ugh.
column_width = max(df[column].astype(str).map(len).max(), len(column))
# set special column widths
narrower_col = ['OS','URL'] #change to fit your workbook
if column in narrower_col: column_width = 10
if column_width >30: column_width = 30
if column == 'IP Address': column_width = 15 #change for your workbook
col_index = df.columns.get_loc(column)
writer.sheets[sheetname].set_column(col_index,col_index,column_width)
return
# Note nothing is returned. Writer.sheets is global.
with pd.ExcelWriter(Output_Wkbk,engine='xlsxwriter') as writer:
# Iterate throuth he series of sheetnames
for tab in tabs:
df1 = pd.read_excel(ar_source, tab).astype(str)
# I need to trim my input
df1.drop(list(df1)[23:],axis='columns', inplace=True, errors='ignore')
try:
# Set spreadsheet focus
df1.to_excel(writer, sheet_name=tab, index = False, na_rep=' ')
# Do something with the spreadsheet - Calling a function
default_col_width(df1, tab, writer)
except:
# Function call failed so just copy tab with no changes
df1.to_excel(writer, sheet_name=tab, index = False,na_rep=' ')
If I use the input file name as the output file name, it fails and erases the original. No need to save or close if you use With... it closes autmatically.
I'd like for the code to run 12345 thru the loop, input it in a worksheet, then start on 54321 and do the same thing except input the dataframe into a new worksheet but in the same workbook. Below is my code.
workbook = xlsxwriter.Workbook('Renewals.xlsx')
groups = ['12345', '54321']
for x in groups:
(Do a bunch of data manipulation and get pandas df called renewals)
writer = pd.ExcelWriter('Renewals.xlsx', engine='xlsxwriter')
worksheet = workbook.add_worksheet(str(x))
renewals.to_excel(writer, sheet_name=str(x))
When this runs, I am left with a workbook with only 1 worksheet (54321).
try something like this:
import pandas as pd
#initialze the excel writer
writer = pd.ExcelWriter('MyFile.xlsx', engine='xlsxwriter')
#store your dataframes in a dict, where the key is the sheet name you want
frames = {'sheetName_1': dataframe1, 'sheetName_2': dataframe2,
'sheetName_3': dataframe3}
#now loop thru and put each on a specific sheet
for sheet, frame in frames.iteritems(): # .use .items for python 3.X
frame.to_excel(writer, sheet_name = sheet)
#critical last step
writer.save()
import pandas as pd
writer = pd.ExcelWriter('Renewals.xlsx', engine='xlsxwriter')
renewals.to_excel(writer, sheet_name=groups[0])
renewals.to_excel(writer, sheet_name=groups[1])
writer.save()
Building on the accepted answer, you can find situations where the sheet name will cause the save to fail if it has invalid characters or is too long. This could happen if you are using grouped values for the sheet name as an example. A helper function could address this and save you some pain.
def clean_sheet_name(sheet):
"""Clean sheet name so that it is a valid Excel sheet name.
Removes characters in []:*?/\ and limits to 30 characters.
Args:
sheet (str): Name to use for sheet.
Returns:
cleaned_sheet (str): Cleaned sheet name.
"""
if sheet in (None, ''):
return sheet
clean_sheet = sheet.translate({ord(i): None for i in '[]:*?/\\'})
if len(clean_sheet) > 30: # Set value you feel is appropriate
clean_sheet = clean_sheet[:30]
return clean_sheet
Then add a call to the helper function before writing to Excel.
for sheet, frame in groups.items():
# Clean sheet name for length and invalid characters
sheet = clean_sheet_name(sheet)
frame.to_excel(writer, sheet_name = sheet, index=False)
writer.save()