On sometimes needing to explicitly copy Python dicts

I recently ran into a Python problem that stumped me for a bit. I was looping through a list of dicts and modifying the dicts slightly, then appending them to a new list. But it wasn’t behaving like I thought it would. Despite my loop being extremely simple and easy to step through mentally, my new list wasn’t always receiving the correct dict. I was befuddled.

The problem was that Python was not making a copy of the dict with each loop. I was modifying the same dict over again in each cycle. Apparently, this is because Python doesn’t inherently implicitly copy objects. I needed to explicitly copy the dict, make my changes, and append the modified copy to the new list.

The working version looked like this:

for field in json_list:
    pl = field["Possible Locations"].split(",")
    for location in pl:
        modified_field = field.copy()
        modified_field["PossibleLocations"] = location.strip()
        output.append(modified_field)

The trick here was to add field.copy() to explicitly copy the dict. Problem solved.

This entry was posted in python. Bookmark the permalink. Both comments and trackbacks are currently closed.

One Comment

  1. Posted August 16, 2024 at 12:46 pm | Permalink

    Worth keeping in mind that you may need copy.deepcopy() in some situations

  • Subscribe to this blog

Skip to toolbar