arcpy.management.DeleteField is not deleting the field?

[HannesZiegler2]

I have the following code (excerpt from script), I intend it to change the field format in a table from a string type to a numerical type. 

gm_reh_sum = r"C:\in_table

gm_reh_sum_mem = arcpy.management.CopyRows(gm_reh_sum, r"in_memory/gm_reh_sum_mem")
 gm_reh_sum_mem = arcpy.management.MakeTableView(gm_reh_sum_mem, "gm_reh_sum_mem")

def change_fldtype(in_table, in_fld, out_fld_type): ##Change field type
   arcpy.management.AddField(in_table, "temp", out_fld_type)
    arcpy.management.CalculateField(in_table, "temp", "[{}]".format(in_fld))
    arcpy.AddMessage([x.name for x in arcpy.ListFields(in_table)]) ##DEBUG##
    arcpy.management.DeleteField(in_table, in_fld)
    arcpy.AddMessage(in_fld)
    arcpy.AddMessage([x.name for x in arcpy.ListFields(in_table)]) ##DEBUG##
    arcpy.management.AddField(in_table, in_fld, out_fld_type)
    arcpy.management.CalculateField(in_table, in_fld, "[temp]")
    arcpy.management.DeleteField(in_table, "temp")

change_fldtype(gm_reh_sum_mem, "Length_Cleaned__Feet_", "DOUBLE")‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍

However, when I run it, for some reason, the delete field in the line arcpy.management.DeleteField(in_table, in_fld) does not delete the field, causing the script to fail when it then tries to add a new field of the new data type with the same name. This is the error that I get (I've also included the debug lines printed before the error message).

Running script WorkSummaryDataImport...
[u'OBJECTID', u'Date_of_Repair__MM_DD_YYYY_', u'Main_Line_Asset_ID', u'Mechanical_Heavy_Sewer_Cleaning___Root_Removal__Yes_No_', u'Length_Cleaned__Feet_', u'Pipe_Burst__Yes_No_', u'CIPP__Yes_No_', u'Open_Cut__Yes_No_', u'Length_Repaired__Feet_', u'Point_Repair_Location__LF_from_US_MH_', u'Point_Repair_Material__PACP_Code_', u'Point_Repair_Greater_than_16_feet_deep___Yes_No_', u'Point_Repair_Length__Feet_', u'Repair_Diameter__inches_', u'Contractor_Comments', u'temp']
Length_Cleaned__Feet_
[u'OBJECTID', u'Date_of_Repair__MM_DD_YYYY_', u'Main_Line_Asset_ID', u'Mechanical_Heavy_Sewer_Cleaning___Root_Removal__Yes_No_', u'Pipe_Burst__Yes_No_', u'CIPP__Yes_No_', u'Open_Cut__Yes_No_', u'Length_Repaired__Feet_', u'Point_Repair_Location__LF_from_US_MH_', u'Point_Repair_Material__PACP_Code_', u'Point_Repair_Greater_than_16_feet_deep___Yes_No_', u'Point_Repair_Length__Feet_', u'Repair_Diameter__inches_', u'Contractor_Comments', u'temp', u'Length_Cleaned__Feet_']
Failed script WorkSummaryDataImport...

Traceback (most recent call last):
 File "C:\Users\zieglerhm\Documents\ArcGIS\Projects\CW2020_IRP_DataProcessing\Script\WorkSummaryDataImport_v0.3.py", line 179, in <module>
 change_fldtype(gm_reh_sum_mem, "Length_Cleaned__Feet_", "DOUBLE")
 File "C:\Users\zieglerhm\Documents\ArcGIS\Projects\CW2020_IRP_DataProcessing\Script\WorkSummaryDataImport_v0.3.py", line 20, in change_fldtype
 arcpy.management.AddField(in_table, in_fld, out_fld_type)
 File "c:\program files (x86)\arcgis\desktop10.3\arcpy\arcpy\management.py", line 3246, in AddField
 raise e
ExecuteError: Failed to execute. Parameters are not valid.
ERROR 000012: Length_Cleaned__Feet_ already exists
Failed to execute (AddField).‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍

Note that before I try to change the field format, I copy the table to memory and then create a table view object on which to operate (this is necessary for later portions of my script not shown here).

Thank you for any help or pointers in the right direction!

[JoshuaBixby · Accepted solution]

I think you have run into a bug.  Using Copy Rows and Make Table View, I can replicate your problem.  Using Copy Features and Make Feature Layer, your code works fine.  So, why does it work with feature layers and not table views?  That is where I think the bug comes into play.

So, seeing there is a bug, what can you do about it?  In this case, the workaround is fairly simple.  Instead of passing the table view, pass the actual name of the in-memory table.  Right now, your code overwrites gm_reh_sum_mem with the result from creating the table view.  I recommend having different variables to hold the in-memory table result and make table view result, and then pass the in-memory table name to your function.

[HannesZiegler2]

Yes, this is what I ended up having to do, thank you!