Reputation: 17601
I am trying to have a python hierarchical data structure with a map and the value will be a tuple. In some cases the tuple will be of length 1. Python intelligently flattens the structure whenever the tuple is of length 1. Observe the example below which can be run in a python interpreter. In the "another_scenario" I expected the length to be 1 but it has drilled one level below and got the underlying steps. This totally screws up my tests because I rely on that being a tuple with the command, function_list, function list.
Question - Why does this happen? How do I ask python to not flatten it?
import os
def run():
my_scenario = {
"scenario_name" :
( # Each scenario is a List of (command, function_list, function_list)
# function_list = one function OR tuple of functions
(
"command1",
(
os.path,
os.path.exists
),
None
),
(
"command2",
(
os.path,
os.path.exists
),
None
)
)
}
another_scenario = {
"scenario_name" :
(
(
"command1",
(
os.path,
os.path.exists
),
None
)
)
}
for name in my_scenario:
print "Full Scenario is %s" % str(my_scenario[name])
print "Length should be 2 -> %s" % len(my_scenario[name])
for name in another_scenario:
print "Full Scenario is %s" % str(another_scenario[name])
print "Length should be 1 -> %s" % len(another_scenario[name]) #Prints 3 as it drills one level down
if __name__ == "__main__":
run()
Upvotes: 0
Views: 370
Reputation: 1122312
You need to add a comma:
another_scenario = {
"scenario_name":
(
(
"command1",
(
os.path,
os.path.exists
),
None
), # <- Note this comma
)
}
to make that a tuple, otherwise it is just an expression. 1-element tuples can only be distinguished from expressions by the presence of a comma:
>>> (1)
1
>>> (1,)
(1,)
>>> type((1))
<type 'int'>
>>> type((1,))
<type 'tuple'>
In fact, it's the comma that defines tuples, not the parenthesis:
>>> 1,
(1,)
>>> 1, 2
(1, 2)
The parenthesis are only needed when you need to define an empty tuple:
>>> ()
()
Upvotes: 2