Making a database with custom commands in python

Question

I'm trying to make a simple, local database using Python where I can set values, get values, etc and I keep getting this error:

#Simple Database
#Functions include Set[name][value]
# Get[name]
# Unset[name]
# Numequalto[value]
# End
# Begin, Rollback, Commit
varlist = []
ops = []
class item:
 def __init__(self,name,value):
 self.name = name
 self.value = value
class db:
 def __init__(self):
 self.varlist = []
 self.ops = []
 def Set(self,nm,val):
 changed = False #Bool for keeping track of update
 for item in varlist: #run through current list
 if item.name == nm: #If the name is found
 item.value = val #update the value
 changed = True #found it
 break #exit if found
 if not changed:
 newitem = item() #Create a new one and append it
 newitem.name = nm
 newitem.value = val
 varlist.append(newitem)
 def Get(key):
 for item in varlist:
 if item.name == key:
 return item.value
 break
 def Unset(key):
 for item in varlist:
 if item.name == key:
 item.value = -1
 break
 def Numequalto(key):
 count = 0
 for item in varlist:
 if item.value == key:
 count+=1
 return count
def main():
 newdb = db()
 varlist=[]
 comm = "" #prime it
 while comm.lower() != "end":
 comm = input("Command: ")
 if comm.lower() == "begin":
 print("----SESSION START---")
 while comm.lower() != "end":
 comm = input("Command: ")
 part = []
 for word in comm.split():
 part.append(word.lower())
 print(part)
 if part[0].lower()=="set":
 newdb.Set(part[1],part[2])
 print(varlist)
 elif part[0].lower()=="get":
 gotten = Get(part[1])
 print(gotten)
 elif part[0].lower()=="unset":
 Unset(part[1])
 elif part[0].lower()=="numequalto":
 numequal = Numequalto(part[1])
 print(numequal)
 print("Finished")
 else:
 print("--ERROR: Must BEGIN--")
if __name__ == "__main__":
 main()

When I run this, and try to create a new item in my list using the command

set a 25

I get this error:

Traceback (most recent call last):
File "/Volumes/CON/LIFE/SimpleDatabase.py", line 81, in <module>
 main()
File "/Volumes/CON/LIFE/SimpleDatabase.py", line 65, in main
 newdb.Set(part[1],part[2])
File "/Volumes/CON/LIFE/SimpleDatabase.py", line 27, in Set
 newitem = item() #Create a new one and append it
UnboundLocalError: local variable 'item' referenced before assignment

Any help would be much appreciated. I'm pretty new to Python

• Waleed Khan: The item class is shadowed by the item in the for-loop. You should use the CamelCase convention for your class names: name your class Item and your variable item.
• James Thiele: Why are you rolling your own DB instead of using the aqlite3 module tha ships with Python?
• cmd: please name your classes with a capital as the first letter, totally threw me off for a second
• abarnert: @JamesThieleHe's not trying to roll his own SQL database, but his own key-value database, so sqlite3 would not be helpful. Of course dbm (or, for 2.x, anydbm) and/or shelve would do the trick. Or, for that matter, a plain old dict, since he's just trying to store his key-value database in memory...
• cmd: reading pep8 should help your coding style

Answer 1

You have a few issues with your code:

1. You are shadowing the class item with a local variable of the same name.
2. You are using varlist instead of self.varlist.
3. Some of your class methods doesn't recieve a self first argument.
4. Also there is a strong convention in python to name classes with a first capital letter.

• abarnert: Whoever downvoted this, why? All of these points are true, except maybe technically #3 (as long as the OP consistently ignores self.varlist and instead uses the global varlist, things will sort of work by accident).
• abarnert: That being said, this isn't an exhaustive list. From a quick glance, besides the blatant errors (e.g., item takes two arguments in its constructor, but instead he's trying to construct it with no arguments and then set the attributes after the fact), there are a whole lot of things that are just very confusing very bad ideas (e.g., having global, local-to-main, and instance-attribute-of-db variables with the same name, using globals as instance attributes, using classic classes...).
• cmd: 2 is not necessarily a problem, it would use the global varlist, but its hard to tell if that is what he meant, since there are 3 different varlists in the program

Answer 2

The line

for item in varlist: 

shadows the class item with a local variable. So that when you get to your item() it thinks you are trying to call the local variable instead of the class. You can tell that your class item is never being constructed because it would fail as you are passing no parameters to the __init__

Also you should really call your class Item. Once I did that I got the constructor error as expected.

Answer 3

Not trying to be implite, just constructive here. I'm concerned that while there are comments questioning the intent to implement your own dictionary, no answer stated this forcefully. I say this only because part of Python (beyond the semantics, language, etc...) is the culture. We speak of things being 'Pythonic' for a reason - part of the value of this language is the culture. There are two aspects here to pay attention to - first, 'Batteries Included' and second, "Don't Reinvent the Wheel". You're reimplimenting the most fundamental composite (oxymoron, I know) data type in Python.

>>> a = {}
>>> a['bob'] = 1
>>> a['frank'] = 2
>>> a
{'frank': 2, 'bob': 1}
>>> del a['frank']
>>> a
{'bob': 1}
>>> del a['bob']
>>> a
{}
>>> a['george'] = 2
>>> b = len([x for x in a.values() if x == 2])
>>> b
1

And there you have it - the pythonic way of handling the functionality you're after.

If you're trying to add functionality or limitations beyond that, you're better off starting from the dict class and extending rather than rolling your own. Since Python is "duck-typed" there's a HUGE benefit to using the existing structure as your basis because it all falls into the same patterns.