How to find maximum values of specific data type in list of tuples and update dictionary with these-CodePudding

If I have the dictionary

max_values = {
    'int': 0,
    'float': 0
}

and I have such a list of tuples

mylist = [(1, 'int'), (3, 'int'), (2, 'float'), (4, 'float')]

How can I find the maximum value of the specific data_type and update the dictionary values?

The dictionary should be

max_values = {
    'int': 3,
    'float': 4
}

How to find maximum values and create new list?

new_list = [(3, 'int'), (4, 'float')]

CodePudding user response：

for value, data_type in mylist:
    max_values[data_type] = max(value, max_values[data_type])

CodePudding user response：

pandas is great for this especially if you have a large dataset:

import pandas as pd

mylist = [(1, 'int'), (3, 'int'), (2, 'float'), (4, 'float')]

a = pd.DataFrame(mylist).groupby(1).max().to_dict()[0]

Or you can use itertools.groupby:

from itertools import groupby
from operator import itemgetter

get_val, get_type = map(itemgetter, (0, 1))
max_values = {k: max(map(get_val, g)) for k, g in groupby(mylist, get_type)}

I used operator.itemgetter as it's more efficient than a lambda expression.

CodePudding user response：

You can first generate the dictionary with checking the maximum value of key for each type . Then using list comprehension you can generate the new list as shown below.

for key, typ in mylist:
    if key > max_values[typ]:
        max_values[typ] = key
print(max_values)
new_list = [(x,y) for y,x in list(max_values.items())]
print(new_list)

This will give outputs for dictionary and list as below.

{'int': 3, 'float': 4}
[(3, 'int'), (4, 'float')]

CodePudding user response：

It can be a one liner:

new_list = [max(mylist, key=lambda item: item[0] if item[1]==tp else 0) for tp in {v for k, v in mylist}]

You can find the types by creating a set

types = {v for _, v in mylist}
# {'int', 'float'}

you can find only the elements with a specific type by:

int_vals = [(val, tp) for val, tp in mylist if tp == 'int']

and you can find the max of the list above by telling max that you're interested in the max of the val part:

max(int_vals, key=lambda item: item[0])

you can combine the last two steps with:

max_int_val = max(mylist, key=lambda item: item[0] if item[1] == 'int')

and then you just need to iterate over the types to create the new list

new_list = []
for tp in types:
    new_list.append(max(mylist key=lambda item: item[0] if item[1] == tp)

which is an expanded version of the one-liner at the top.

If you also need the max_values dict, you can get it by:

max_values = {v: k for k, v in new_list}