Search code examples
pythondefaultdict

Python multi level default dict

I am making inverted index.For this purpose I am taking values from a file.Each value of a file is in form:

document_Id'\t'term_Id'\t'pos_1'\t'pos_2...'\t'pos_n

This is a forward index representation.I want to convert it into inverted index which should look like

term_Id'\t'"doc_Id:pos1,pos2...posn""doc_Id:pos1,pos2...posn"

For that purpose I am using default dict of list type.This is my function:

nestedDict = defaultdict(lambda:defaultdict(list))

def getInfo(line):
    global nestedDict
    tokens = re.split(r'\t+',line)
    docInfo = int(tokens[0]) #Set document Id
    termId = int(tokens[1]) #Set Term Id
    currentPosition = int(tokens[2])
    nestedDict[str(termId)][str(docInfo)] = str(currentPosition)        
    if len(tokens) > 3 :
        for i in range(3,len(tokens)):
            position = int(tokens[i])-currentPosition
            currentPosition = currentPosition + position
            nestedDict[str(termId)][str(docInfo)].append(currentPosition)

It is giving me an error:Str has no method .append. I am new to python.Any help would be highly appreciated.

Solution

  • Your nested defaultdict makes nestedDict[...][...] be a list, but then you assign a string to it. I don't think you need that assignment anyway: why not just let the loop handle all of the positions?