Deleting rows from an access table based on multiple conditions in python along with a date field

Deleting rows from an access table based on multiple conditions in python along with a date field



I am trying to delete rows from an access database table, based on two columns one which is released_by and released_date.


Sudo code

WHERE released_by = '27' and released_date would change based on the day of the month.



Released_Date if day ==1 then delete previous month's data else current months data.


import pandas as pd
import numpy as np
import datetime
from datetime import date
from datetime import datetime, timedelta
from dateutil.relativedelta import relativedelta
import pyodbc
import calendar
import xlrd
import defusedxml
from defusedxml.common import EntitiesForbidden
from xlrd import open_workbook
defusedxml.defuse_stdlib()

# connecting to access database
conn = pyodbc.connect(r'Driver=Microsoft Access Driver (*.mdb, *.accdb);DBQ=C:UsersDesktopInsights.accdb;')
insights_dbcurs = conn.cursor()

select_auto_releases = "SELECT * FROM RELEASES WHERE RELEASED_BY = '27'"
autoreleases = insights_dbcurs.execute(select_auto_releases).fetchall()
#display(autoreleases)

for row in autoreleases:
previousmonth = datetime.now() - relativedelta(months=1)
previousmonth = previousmonth.strftime("%m - %Y")
currentmonth = datetime.now()
currentmonth = currentmonth.strftime("%m - %Y")
if ((row.autoreleases['RELEASED_DATE']).strftime) ==1:
try:
delete = 'DELETE * FROM RELEASES WHERE RELEASED_DATE = ? and RELEASED_BY = ?'
insights_dbcurs.execute(delete,'RELEASED_DATE':currentmonth,'RELEASED_BY':'27')
except:
delete = 'DELETE * FROM RELEASES WHERE RELEASED_DATE = ? and RELEASED_BY = ?'
insights_dbcurs.execute(delete,'RELEASED_DATE':previousmonth,'RELEASED_BY':'27')

insights_dbcurs.close()
conn.commit()
conn.close()



It does not work. I get:



AttributeError: 'pyodbc.Row' object has no attribute 'autoreleases'



if I remove Row.autorelease, I get


Row.autorelease



sql has 0 attributes, 2 being passed.



How do i avoid this. Any suggestions would be really helpful.






Try if ((row.RELEASED_DATE).strftime) ==1:

– Gord Thompson
Sep 14 '18 at 18:08


if ((row.RELEASED_DATE).strftime) ==1:






@GordThompson Thanks i tried. i noticed while running the code now is that RELEASED_DATE is being pulled in as a string. 'AttributeError: 'str' object has no attribute 'strptime' or strftime '. when i tried to convert it to a date format. also i tried .((row.RELEASED_DATE).astype(date).strftime ==1it did not work as well. The above one worked apart from this new fallout.

– Ren Lyke
Sep 14 '18 at 18:30



'AttributeError: 'str' object has no attribute 'strptime' or strftime '


((row.RELEASED_DATE).astype(date).strftime ==1






If row.RELEASED_DATE is a string then either use .strptime(...) to parse it and then .strftime("%m - %Y") to re-format it, or just use string manipulation to re-arrange the substring parts to match .strftime("%m - %Y")

– Gord Thompson
Sep 14 '18 at 19:01


row.RELEASED_DATE


.strptime(...)


.strftime("%m - %Y")


.strftime("%m - %Y")






since i am using the current date to check which month's data to be deleted i changed it to this if datetime.now().strftime('%d') !=1: ` try:` delete = 'DELETE * FROM RELEASES WHERE RELEASED_DATE = ? and RELEASED_BY = ?' insights_dbcurs.execute(delete,'RELEASED_DATE':currentmonth,'RELEASED_BY':'27') ` however i get this SystemError: <class 'pyodbc.ProgrammingError'> returned a result with an error set. Error points near the execute statement.

– Ren Lyke
Sep 14 '18 at 19:19



if datetime.now().strftime('%d') !=1:


delete = 'DELETE * FROM RELEASES WHERE RELEASED_DATE = ? and RELEASED_BY = ?'


insights_dbcurs.execute(delete,'RELEASED_DATE':currentmonth,'RELEASED_BY':'27')


SystemError: <class 'pyodbc.ProgrammingError'> returned a result with an error set




1 Answer
1



Consider re-factoring your code for several issues:



try/except is usually used for handling runtime exceptions and errors, not application logic. Use the if/else for these situations.


try/except


if/else



Remove the unneeded for loop as you are not deleting row by row since no unique identifier is being passed into DELETE statements. In actuality, you are running the deletion process across whole table if the logic passes that current row not filtered to that one row.


for


DELETE



Use pure SQL with one DELETE statement and avoid the Python handling of time elements which as you see maintains conversion issues from the database. MS Access SQL does have date functions such as Date() (current date) and DatePart() for month/day extractions. Additionally, MS Access can run complex subquery logic in DELETE.


DELETE


Date()


DatePart()


DELETE



SQL



Using IN subquery for both date conditions. NOTE: ID should be replaced with unique identifier of table. Check SELECT version of query prior to DELETE version.


IN


ID


SELECT


DELETE


DELETE FROM RELEASES r
WHERE r.ID IN
(SELECT sub.ID
FROM RELEASES sub
WHERE sub.RELEASED_BY = ?
AND (
(
DatePart('m', sub.RELEASED_DATE) = DatePart('m', Date()) - 1
AND DatePart('yyyy', sub.RELEASED_DATE) = DatePart('yyyy', Date())
AND DatePart('d', sub.RELEASED_DATE) = 1
)
OR
(
DatePart('m', sub.RELEASED_DATE) = DatePart('m', Date())
AND DatePart('yyyy', sub.RELEASED_DATE) = DatePart('yyyy', Date())
AND DatePart('d', sub.RELEASED_DATE) > 1
)
)
)



Python



Passing parameters with dates handled in Access engine.


conn = pyodbc.connect(r'...')
insights_dbcurs = conn.cursor()

sql = """DELETE FROM RELEASES r
WHERE r.ID IN
(SELECT sub.ID
FROM RELEASES sub
WHERE sub.RELEASED_BY = ?
AND (
(
DatePart('m', sub.RELEASED_DATE) = DatePart('m', Date()) - 1
AND DatePart('yyyy', sub.RELEASED_DATE) = DatePart('yyyy', Date())
AND DatePart('d', sub.RELEASED_DATE) = 1
)
OR
(
DatePart('m', sub.RELEASED_DATE) = DatePart('m', Date())
AND DatePart('yyyy', sub.RELEASED_DATE) = DatePart('yyyy', Date())
AND DatePart('d', sub.RELEASED_DATE) > 1
)
)
)"""

insights_dbcurs.execute(sql, ('27',))
conn.commit()
conn.close()






Thanks for the modification. I have tried the same and continue to get an error at the execute section. ProgrammingError: ('42000', "[42000] [Microsoft][ODBC Microsoft Access Driver] Invalid use of '.', '!', or '()'. in query expression 'SUB.RELEASED_BY = Pa_RaM000n AND n (n (DatePart('m',SUB.RELEASED_DATE) = DatePart('m'.Date())-1n AND DatePart('d',SUB.RELEASED'. (-3100) (SQLExecDirectW)").

– Ren Lyke
Sep 14 '18 at 20:03


ProgrammingError: ('42000', "[42000] [Microsoft][ODBC Microsoft Access Driver] Invalid use of '.', '!', or '()'. in query expression 'SUB.RELEASED_BY = Pa_RaM000n AND n (n (DatePart('m',SUB.RELEASED_DATE) = DatePart('m'.Date())-1n AND DatePart('d',SUB.RELEASED'. (-3100) (SQLExecDirectW)")






Seems like a missing parenthesis issue but I cannot see it. I just copied this query into MS Access and it saved without issue. Access does NOT save queries with syntax errors. Did you make any changes from this post anywhere in Python? I am seeing Pa_RaM000. Is that a parameter?

– Parfait
Sep 14 '18 at 20:43






This suggestion might be especially helpful in this case because Access tends to be extra accommodating when it comes to type juggling. For example, Access will automatically convert a date-like string passed in place of a true date (e.g., DatePart("m", "2018-09-12") works), while Python is much stricter about such things (e.g., d.month works if d is a date but not if d is a str).

– Gord Thompson
Sep 14 '18 at 20:48


DatePart("m", "2018-09-12")


d.month


d


date


d


str






@Parfait - Pa_RaM000 is the internal representation of a query PaRaMeter when Jet/ACE is preparing a statement. It can also be found in Jet/ACE SHOWPLAN output.

– Gord Thompson
Sep 14 '18 at 20:52


Pa_RaM000






i checked once again there is not missing parenthesis. this Pa_RaM000 is coming when i use '27' at the end. I have not added anything additional. If i use '?' then i get the error SQL has 0 parameters 2 parameters being passed.

– Ren Lyke
Sep 14 '18 at 21:29


SQL has 0 parameters 2 parameters being passed.



Thanks for contributing an answer to Stack Overflow!



But avoid



To learn more, see our tips on writing great answers.



Required, but never shown



Required, but never shown




By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Popular posts from this blog

𛂒𛀶,𛀽𛀑𛂀𛃧𛂓𛀙𛃆𛃑𛃷𛂟𛁡𛀢𛀟𛁤𛂽𛁕𛁪𛂟𛂯,𛁞𛂧𛀴𛁄𛁠𛁼𛂿𛀤 𛂘,𛁺𛂾𛃭𛃭𛃵𛀺,𛂣𛃍𛂖𛃶 𛀸𛃀𛂖𛁶𛁏𛁚 𛂢𛂞 𛁰𛂆𛀔,𛁸𛀽𛁓𛃋𛂇𛃧𛀧𛃣𛂐𛃇,𛂂𛃻𛃲𛁬𛃞𛀧𛃃𛀅 𛂭𛁠𛁡𛃇𛀷𛃓𛁥,𛁙𛁘𛁞𛃸𛁸𛃣𛁜,𛂛,𛃿,𛁯𛂘𛂌𛃛𛁱𛃌𛂈𛂇 𛁊𛃲,𛀕𛃴𛀜 𛀶𛂆𛀶𛃟𛂉𛀣,𛂐𛁞𛁾 𛁷𛂑𛁳𛂯𛀬𛃅,𛃶𛁼

Crossroads (UK TV series)

ữḛḳṊẴ ẋ,Ẩṙ,ỹḛẪẠứụỿṞṦ,Ṉẍừ,ứ Ị,Ḵ,ṏ ṇỪḎḰṰọửḊ ṾḨḮữẑỶṑỗḮṣṉẃ Ữẩụ,ṓ,ḹẕḪḫỞṿḭ ỒṱṨẁṋṜ ḅẈ ṉ ứṀḱṑỒḵ,ḏ,ḊḖỹẊ Ẻḷổ,ṥ ẔḲẪụḣể Ṱ ḭỏựẶ Ồ Ṩ,ẂḿṡḾồ ỗṗṡịṞẤḵṽẃ ṸḒẄẘ,ủẞẵṦṟầṓế