通过ODBC将熊猫输出连接到Excel工作表

Question

The query below does the following: Connect to a SQL Server db via ODBC, run a SQL script (many queries separated by ";"), create two dataframes for two specific query results, and then export them to two tabs within an excel workbook:

import pyodbc
import pandas as pd
import time

name= 'ouput' + str(time.strftime("%Y-%b-%d__%H_%M_%S",time.localtime()))
print ("Connecting via ODBC")

conn = pyodbc.connect('DSN=Server DB Prod', autocommit=True)

print ("Connected!\n")

inputdir = 'H:\\Queries\\ADS'

#for script in os.listdir(inputdir):
with open(inputdir+'\\' + 'query' +'.sql','r') as inserts:
    sqlScript = inserts.read()
    for statement in sqlScript.split(';'):
        with conn.cursor() as cur:
            cur.execute(statement)

query1="Select * from #leadership"
data1=pd.read_sql_query(query1, conn).sort_values(['channel','terr_code'], ascending=[0,1]).reset_index(drop=True)
#print(data1.head(n=100))
query2="Select * from #ml"
data2=pd.read_sql_query(query2, conn).sort_values(['channel','terr_code','client_name'], ascending=[0,1,1]).reset_index(drop=True)

print('query finished')
conn.close()

writer = pd.ExcelWriter(name+ '.xlsx')
data1.to_excel(writer,'Leadership Summary')
data2.to_excel(writer,'ML Detail')
writer.save()

print("Results were succesfully exported")

However, I am hoping to instead be able to connect to an existing excel file via ODBC in order to update the tabs of my workbook dynamically and not lose the formatting and graphs- to allow real automation. Any other solution that would allow the same will certainly work.

Background: I am trying to automate a process where I run a query in SQL Server (via Python), and have the output update the tabs of an existing excel worksheet- I was thinking by connecting via ODBC. That worksheet has specific formatting, and formulas+graphs built based on the data.

Note: I dont have write permission, only read, so I cant connect a "final" SQL table to excel via ODBC. I am doing some additional data blending from other sources in Python (not shown), so connecting a SQL query to excel via ODBC will not work.

Any help is greatly appreciated.

Answer 1

The best strategy is to access the Excel workbook and use its facilities instead of external tools in order to leave all other objects untouched. Hence, consider the win32com client where you can then access the Excel object library such as its CopyFromRecordset method.

And instead of pyodbc as database API, use Window's ADODB API which can use ODBC connections. Also, there is no need for pandas as worksheet objects are used to hold data. NOTE: this solution only works on Windows machines.

import win32com.client as win32

try:
    # INITIALIZE OBJECTS
    xlapp = win32.gencache.EnsureDispatch('Excel.Application')
    ado_conn = win32.gencache.EnsureDispatch('ADODB.Connection')
    ado_rst = win32.gencache.EnsureDispatch('ADODB.Recordset')

    # OPEN CONNECTION
    ado_conn.Open('DSN=Server DB Prod')

    # RUN QUERIES
    with open(inputdir+'\\' + 'query' +'.sql','r') as inserts:
        sqlScript = inserts.read()
        for statement in sqlScript.split(';'):
            ado_conn.Execute(statement)

    # OPEN WORKBOOK AND UPDATE SHEETS
    xlwb = xlapp.Workbooks.Open(r'C:\Full\Path\To\Workbook.xlsx')

    ls = xlwb.Worksheets('Leadership Summary') 
    ls.Cells.ClearContents()
    ado_rst.Open("Select * from #leadership", ado_conn)
    for i in range(ado_rst.Fields.Count):
       ls.Cells(1, i+1).Value = ado_rst.Fields(i).Name     # COLUMNS
    ls.Range("A2").CopyFromRecordset(ado_rst)              # DATA ROWS
    ado_rst.Close()

    ml = xlwb.Worksheets('ML Detail')
    ml.Cells.ClearContents()
    ado_rst.Open("Select * from #ml", ado_conn)         
    for i in range(ado_rst.Fields.Count): 
       ml.Cells(1, i+1).Value = ado_rst.Fields(i).Name     # COLUMNS
    ml.Range("A2").CopyFromRecordset(ado_rst)              # DATA ROWS
    ado_rst.Close()

    ado_conn.Close()
    xlapp.Visible = True        # OPENS WORKBOOK WITH ABOVE CHANGES TO SCREEN

except Exception as e:
    print(e)

finally:
    # RELEASE RESOURCES
    ls = None; ml = None
    ado_rst = None; ado_conn = None
    xlwb = None; xlapp = None

Hopefully, this debunks any who think VBA (also a COM-interfaced language) is the only coding language for Excel!