"Append" value to existing dictionary key - is there a better way?

Hello everyone,

I am working my way through the Data Scientist Career Path and I am currently on the first non step-by-step project which I am working locally using Jupyter Notebooks:

Project: Hurricane Analysis

I spent almost half the day for completing task #2 and task #3, but I finally succeed to get the required results and I have to say, it felt fulfilling!

Despite that, I believe that the way I approach the problem might not be the most efficient, although after a lot of googling it seems like the only way that achieves the desired result! Thus, I would like to share my code so far and ask if there is an alternative to that.

I have uploaded my code on my GitHub page, as I am working locally for this project.

I would appreciate any advice and feedback!


For step 2 you can use list comprehension (5. Data Structures β€” Python 3.10.2 documentation) and conditional expressions (6. Expressions β€” Python 3.10.2 documentation):

updated_damages = [
  x if x == 'Damages not recorded' else float(x[:-1]) * conversion[x[-1]] 
  for x in damages

Step 3 is more complicated. As start I create a list of dictionaries using zip (Built-in Functions β€” Python 3.10.2 documentation) and list comprehension:

# I think there is more cleaner way to do this
non_indexed_list = [
      { "Name": x[0], "Month": x[1], "Year": x[2], "Max Sustained Wind": x[3], "Areas Affected": x[4], "Damage": x[5], "Deaths": x[6] } 
      for x in zip(names, months, years, max_sustained_winds, areas_affected, updated_damages, deaths)

Next wrote two helper functions:

# For unique fields. Returns { field: dict }
def key_by(data: list, field: str) -> dict:
  return { x[field]: x for x in data }

# For grouping by non-unique field. Returns { field: [dict, dict, ...]  }
def group_by(data: list, field: str) -> dict:
  grouped = {}
  for r in data:
    key = r[field]
    if key not in grouped:
      grouped[key] = []
  return grouped

And then simply called helpers with non_indexed_list:

hurricanes = key_by(non_indexed_list, "Name")
hurricanes_by_year = group_by(non_indexed_list, "Year")

Note that values of hurricanes_by_year have to be lists even in there is only one hurricane in this year.

Hello @9509706156,

Thank you for taking the time to help!

You are a live example that is one step to just achieve the required result, and another to write efficient code alltogether!

I would never have thought of your so succint solution for step 2! I won’t even comment for the 3rd :joy: !


Two more solutions to generate non_indexed_list:

With columns and double zip:

columns = [ "Name", "Month", "Year", "Max Sustained Wind", "Areas Affected", "Damage", "Deaths" ]
non_indexed_list = [
      dict(zip(columns, x))
      for x in zip(names, months, years, max_sustained_winds, areas_affected, updated_damages, deaths) 
      # The sequence of arrays in zip must be the same as names in columns array

And another one with dict-table:

table = {
  "Name": names,
  "Month": months,
  "Year": years,
  "Max Sustained Wind": max_sustained_winds,
  "Areas Affected": areas_affected,
  "Damage": updated_damages,
  "Deaths": deaths,
non_indexed_list = [
      dict(zip(table.keys(), x))
      for x in zip(*table.values())
1 Like