Question

I have a dataset in Pytables, which looks something like

class myData(IsDescription):
    date = StringCol(16)
    item = Int32Col()

I have multiple items per same date, for example:

'2010-01-01', 5
'2010-01-01', 6
'2010-01-02', 7
'2010-01-02', 8

Is there a way to iterate over unique dates and then over items in the date? I mean something like

for date in DATE
    print date
     for ITEM
         print item

Answer 1

I am not familiar with the inner-workings of Pytables (so this may not be in-line with what you are looking for), but the groupby function in the itertools module is very useful in these types of situations (note the sorting step below – this is important in this case in order to get groupby to group all items with the same date. See here for more info.):

In [1]: from itertools import groupby

In [2]: from operator import attrgetter

In [3]: class myData(object):
    def __init__(self, date, item):
        self.date = date
        self.item = item
   ...:         

In [4]: l = [myData('2012-01-01', 'thing'), myData('2012-01-01', 'another thing'), myData('2013-01-01', 'and another')]

In [5]: l_sorted = sorted(l, key=attrgetter('date'))

In [6]: for date, my_objects in groupby(l_sorted, key=attrgetter('date')):
   ...:     print date
   ...:     for obj in my_objects:
   ...:         print obj.item
   ...:         
2012-01-01
thing
another thing
2013-01-01
and another

The basic pattern here is to get a list/container that holds the objects you want to group. You then sort that list based by the attribute by which we will later be grouping (in this case, date). You then pass that sorted list to the groupby function, which will emit two values in each iteration – a key that represents the value by which you grouped (so here it will be the date of each group) and a group iterator that contains all of your objects that share that same date key. You can then iterate through that group, pulling out the item attribute of each object.