Iterate over data to build dictionary

[dshimer]

I have a file whose structure in strictly generic terms is similar to the following.

keyname first
keyword1 1.1
keyword2 1.2
keyword3 1.3
keyname second
keyword1 2.1
keyword2 2.2
keyword3 2.3
keyname third
keyword1 3.1
keyword2 3.2
keyword3 3.3

keyname value is always going to contain the name of a data set identified by keywords and values. The keywords are always the same for each keyname and contain data which is unique to the keyname. In the simplest terms if I read it like

>>> f=open('/tmp/test.txt','r')
>>> d=f.readlines()
>>> for l in d:
... 	print l.split()
... 
['keyname', 'first']
['keyword1', '1.1']
['keyword2', '1.2']
['keyword3', '1.3']
['keyname', 'second']
['keyword1', '2.1']
['keyword2', '2.2']
['keyword3', '2.3']
['keyname', 'third']
['keyword1', '3.1']
['keyword2', '3.2']
['keyword3', '3.3']

What I would like to do is build a dictionary in which each keyname has a value which is another dictionary made up of keywords and values. For example if I were to manually build it the dictionary would look like

dict={'first':{'keyword1':1.1,'keyword2':1.2,'keyword3':1.3},'second':{'keyword1':2.1,'keyword2':2.2,'keyword3':2.3},'third':{'keyword1':3.1,'keyword2':3.2,'keyword3':3.3}}

allowing for access to whole keys, or individual data values like

>>> dict['second']
{'keyword3': 2.2999999999999998, 'keyword2': 2.2000000000000002, 'keyword1': 2.1000000000000001}
>>> dict['second']['keyword2']
2.2000000000000002

This is ripe for iterating over the data and adding as I go, if it were a list I would append, but I don't use dictionaries very often and don't know how to add/append/insert data. What is the best way to do this?

[bartonc]

I have a file whose structure in strictly generic terms is similar to the following.

keyname first
keyword1 1.1
keyword2 1.2
keyword3 1.3
keyname second
keyword1 2.1
keyword2 2.2
keyword3 2.3
keyname third
keyword1 3.1
keyword2 3.2
keyword3 3.3

keyname value is always going to contain the name of a data set identified by keywords and values. The keywords are always the same for each keyname and contain data which is unique to the keyname. In the simplest terms if I read it like

>>> f=open('/tmp/test.txt','r')
>>> d=f.readlines()
>>> for l in d:
... 	print l.split()
... 
['keyname', 'first']
['keyword1', '1.1']
['keyword2', '1.2']
['keyword3', '1.3']
['keyname', 'second']
['keyword1', '2.1']
['keyword2', '2.2']
['keyword3', '2.3']
['keyname', 'third']
['keyword1', '3.1']
['keyword2', '3.2']
['keyword3', '3.3']

What I would like to do is build a dictionary in which each keyname has a value which is another dictionary made up of keywords and values. For example if I were to manually build it the dictionary would look like

dict={'first':{'keyword1':1.1,'keyword2':1.2,'keyword3':1.3},'second':{'keyword1':2.1,'keyword2':2.2,'keyword3':2.3},'third':{'keyword1':3.1,'keyword2':3.2,'keyword3':3.3}}

allowing for access to whole keys, or individual data values like

>>> dict['second']
{'keyword3': 2.2999999999999998, 'keyword2': 2.2000000000000002, 'keyword1': 2.1000000000000001}
>>> dict['second']['keyword2']
2.2000000000000002

This is ripe for iterating over the data and adding as I go, if it were a list I would append, but I don't use dictionaries very often and don't know how to add/append/insert data. What is the best way to do this?

It's simple:

aDict[newKeyName] = newValue

Scary, huh?

[dshimer]

Good grief, I knew it had to be easy. I would be sorry for taking your time instead of digging through a book, but I guess it will be a good reference in case anybody else missed it. Python data types are just too cool and I'm finding that I like dictionaries more than I thought I would when I first started studying them.

Thanks

[bartonc]

Good grief, I knew it had to be easy. I would be sorry for taking your time instead of digging through a book, but I guess it will be a good reference in case anybody else missed it. Python data types are just too cool and I'm finding that I like dictionaries more than I thought I would when I first started studying them.

Thanks

Any time, D. It's really no trouble (and you contribute so much that the simplest to the toughest questions are yours for the asking). And as you say, it could help someone else along the way.

[bvdet]

Good grief, I knew it had to be easy. I would be sorry for taking your time instead of digging through a book, but I guess it will be a good reference in case anybody else missed it. Python data types are just too cool and I'm finding that I like dictionaries more than I thought I would when I first started studying them.

Thanks

I like dictionaries also. I hope this helps you:

data = open(fn).read()
dataLst = [i.strip() for i in data.split('keyname') if i != '']
dd = {}
for item in dataLst:
    itemLst = item.split('\n')
    dd[itemLst[0]] = dict(zip([i.split()[0] for i in itemLst[1:]], [j.split()[1] for j in itemLst[1:]]))

for key in dd:
    print '%s = %s' % (key, dd[key])

'''
>>> second = {'keyword3': '2.3', 'keyword2': '2.2', 'keyword1': '2.1'}
third = {'keyword3': '3.3', 'keyword2': '3.2', 'keyword1': '3.1'}
first = {'keyword3': '1.3', 'keyword2': '1.2', 'keyword1': '1.1'}
>>>
'''

I also apreciate your contributions. :)