Min/Max Values Across Multiple Sublists within a Dictionary - python-3.x

I'm working on a program that will read a .txt file of NBA team names with fives statistics for each team into a dictionary. I was able to read the file into a dictionary with the correct key-value pairs, but I can't figure out how to return a table of minimum, maximum, and average values for each statistic across all teams. I've looked at other questions on the site, but I can't find anything pertaining to a key with multiple values per entry in the dictionary. Here's what the dictionary looks like for a few of the entries:
{'Golden State Warriors':[113.5, 107.5, 43.5, 0.503, 8.0], 'Houston Rockets':[112.4, 103.9, 43.5, 0.46, 8.5], ... : ...}
I need to make a table that displays the min, max, and average of each statistic across all entries:
PPG PPAG RPG FG SPG
MIN
MAX
AVG
I can do this kind of stuff with the statistics in the form of a list, but whenever I try to write lists into a dictionary, I get a TypeError. Would greatly appreciate suggestions, I've been stuck on this for hours.
Also, IMPORTANT NOTE: I cannot use lambda for this task, I am working on this for a project and my options are for loops, while loops, and the basic list functions and dictionary methods.

The following code should do the computation of the stas:
from statistics import mean
lists = list(dic.values())
output = {'MIN':[], 'MAX':[], 'AVG':[]}
for i in range(0, len(lists[0])):
output['MIN'].append(min(col[i] for col in lists))
output['MAX'].append(max(col[i] for col in lists))
output['AVG'].append(mean(col[i] for col in lists))
To print the results as a table, you may do the following:
print("\tPPG\tPPAG\tRPG\tFG\tSPG")
for stats in output:
print(stats, end='\t')
for i in range(0, cols):
print(round(output[stats][i], 3), end='\t')
print()
Hope it helps.

As long as all the data for each team is in the same order in the dictionary this will work well for your purpose. I started by forming a list containing the dictionary values which are also lists. Then you can zip the sub lists together, this has the handy feature of grouping all the same indices in the lists together. Now we can call max, min, and mean on each zip'd stat! If you can not import the 'statistics' module a list comp would also be an easy one liner.
from statistics import mean
stats = {'Golden State Warriors':[113.5, 107.5, 43.5, 0.503, 8.0], 'Houston Rockets':[112.4, 103.9, 43.5, 0.46, 8.5]}
values = [*zip(*stats.values())]
for value in values:
print("max: {}\nmin: {}\nmean: {:0.3f} \n".format(max(value), min(value), mean(value)))
EDIT: take note that if you try to zip lists together that are not the same length it will zip by the shortest list. This behavior can be corrected by using the itertools izip_longest and providing a default value.

Related

Nested loops altering rows in pandas - Avoiding "A value is trying to be set on a copy of a slice from a DataFrame"

Summary
I am trying to loop through a pandas dataframe, and to run a secondary loop at each iteration. The secondary loop calculates something that I want to append into the original dataframe, so that when the primary loop advances, some of the rows are recalculated based on the changed values. (For those interested, this is a simple advective model of carbon accumulation in soils. When a new layer of soil is deposited, mixing processes penetrate into older layers and transform their properties to a set depth. Thus, each layer deposited changes those below it incrementally, until a former layer lies below the mixing depth.)
I have produced an example of how I want this to work, however it is generating the common error message:
A value is trying to be set on a copy of a slice from a DataFrame
See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
self._setitem_single_block(indexer, value, name)
I have looked into the linked information in the error message as well as myriad posts on this forum, but none get into the continual looping through a changed dataframe.
What I've tried, and some possible solutions
Below is some example code. This code works more or less as well as I want it to. But it produces the warning. Should I:
Suppress the warning and continue working with this architecture? In this case, am I asking for trouble with un-reproducible results?
Try a different architecture altogether, like a numpy array from the original dataframe?
Try df.append() or df.copy() to avoid the warning?
I have tried `df.copy()' to no avail - the warning was still thrown.
Example code:
import pandas as pd
a = pd.DataFrame(
{
'a':[x/2 for x in range(1,11)],
'b':['hot dog', 'slider', 'watermelon', 'funnel cake', 'cotton candy', 'lemonade', 'fried oreo', 'ice cream', 'corn', 'sausage'],
'c':['meat', 'meat', 'vegan', 'vegan', 'vegan', 'vegan', 'dairy','dairy', 'vegan', 'meat']
}
)
print(a)
z = [x/(x+2) for x in range(1,5)]
print(z)
#Primary loop through rows of the main dataframe
for ind, row in a.iterrows():
#Pull out a chunk of the dataframe. This is the portion of the dataframe that will be modified. What is below
#this is already modified and locked into the geological record. What is above has not yet been deposited.
b = a.iloc[ind:(ind+len(z)), :]
#Define the size of the secondary loop. Taking the minimum avoids the model mixing below the boundary layer (key error)
loop = min([len(z), len(b)])
#Now loop through the sub-dataframe and change accordingly.
for fraction in range(loop):
b['a'].iloc[fraction] = b['a'].iloc[fraction]*z[fraction]
#Append the original dataframe with new data:
a.iloc[ind:(ind+loop), :] = b
#Try df.copy(), but still throws warning!
#a.iloc[ind:(ind+loop), :] = b.copy()
print(a)

Sum all counts when their fuzz.WRatio > 90 otherwise leave intact

What I want to do was actually group by all similar strings in one columns and sum their
corresponding counts if there are similarity, otherwise, leave them.
A little similar to this post. Unfortunately I have not been able to apply this to my case:
How to group Pandas data frame by column with regex match
Unfortunately, I ended up with the following steps:
I wrote a function to print out all the fuzz.Wratio for each row of string,
when each row does a linear search from the top to check if there are other similar
strings in the rest of the rows. If the WRatio > 90, I would like to sum these row's
corresponding counts. Otherwise, leave them there.
I created a test data looking like this:
test_data=pd.DataFrame({
'name':['Apple.Inc.','apple.inc','APPLE.INC','OMEGA'],
'count':[4,3,2,6]
})
So what I want to do is make the result as a dataframe like:
result=pd.Dataframe({
'Nname':['Apple.Inc.','OMEGA'],
'Ncount':[9,6]
})
My function so far only gave me the fuzz ratio for each row,
and to my understanding is that,
each row compares to itself three times( here we have four rows).
So My function output would look like:
pd.Dataframe({
'Nname':['Apple.Inc.','Apple.Inc.','Apple.Inc.','apple.inc',\
'apple.inc','apple.inc'],
'Ncount':[4,4,4,3,3,3],
'FRatio': [100,100,100,100,100,100] })
This is just one portion of the whole output from the function I wrote with this test data.
And the last row "OMEGA" would give me a fuzz ratio about 18.
My function is like this:
def checkDupTitle2(data):
Nname=[]
Ncount=[]
f_ratio=[]
for i in range(0, len(data)):
current=0
count=0
space=0
for space in range(0, len(data)-1-current):
ratio=fuzz.WRatio(str(data.loc[i]['name']).strip(), \
str(data.loc[current+space]['name']).strip())
Nname.append(str(data.loc[i]['name']).strip())
Ncount.append(str(data.loc[i]['count']).strip())
f_ratio.append(ratio)
df=pd.DataFrame({
'Nname': Nname,
'Ncount': Ncount,
'FRatio': f_ratio
})
return df
So after running this function and get the output,
I tried to get what I eventually want.
here I tried group by on the df created above:
output.groupby(output.FRatio>90).sum()
But this way, I still need a "name" in my dataframe,
how can I decide on which names for this total counts, say, 9 here.
"Apple.Inc" or "apple.inc" or "APPLE.INC"?
Or, did I make it too complex?
Is there a way to group by "name" at the very first and treat "Apple.Inc.", "apple.inc" and "APPLE.INC" all the same, then my problem has solved. I have stump quite a while. Any helps would be highly
appreciated! Thanks!
The following code is using my library RapidFuzz instead of FuzzyWuzzy since it is faster and it has a process method extractIndices which does help here. This solution is quite a bit faster, but since I do not work with pandas regulary I am sure there are still some things that could be improved :)
import pandas as pd
from rapidfuzz import process, utils
def checkDupTitle(data):
values = data.values.tolist()
companies = [company for company, _ in values]
pcompanies = [utils.default_process(company) for company in companies]
counts = [count for _, count in values]
results = []
while companies:
company = companies.pop(0)
pcompany = pcompanies.pop(0)
count = counts.pop(0)
duplicates = process.extractIndices(
pcompany, pcompanies,
processor=None, score_cutoff=90, limit=None)
for (i, _) in sorted(duplicates, reverse=True):
count += counts.pop(i)
del pcompanies[i]
del companies[i]
results.append([company, count])
return pd.DataFrame(results, columns=['Nname','Ncount'])
test_data=pd.DataFrame({
'name':['Apple.Inc.','apple.inc','APPLE.INC','OMEGA'],
'count':[4,3,2,6]
})
checkDupTitle(test_data)
The result is
pd.Dataframe({
'Nname':['Apple.Inc.','OMEGA'],
'Ncount':[9,6]
})

Slow loop aggregating rows and columns

I have a DataFrame with a column named 'UserNbr' and a column named 'Spclty', which is composed of elements like this:
[['104', '2010-01-31'], ['215', '2014-11-21'], ['352', '2016-07-13']]
where there can be 0 or more elements in the list.
Some UserNbr keys appear in multiple rows, and I wish to collapse each such group into a single row such that 'Spclty' contains all the unique dicts like those in the list shown above.
To save overhead on appending to a DataFrame, I'm appending each output row to a list, instead to the DataFrame.
My code is working, but it's taking hours to run on 0.7M rows of input. (Actually, I've never been able to keep my laptop open long enough for it to finish executing.)
Is there a better way to aggregate into a structure like this, maybe using a library that provides more data reshaping options instead looping over UserNbr? (In R, I'd use the data.table and dplyr libraries.)
# loop over all UserNbr:
# consolidate specialty fields into dict-like sets (to remove redundant codes);
# output one row per user to new data frame
out_rows = list()
spcltycol = df_tmp.column.get_loc('Spclty')
all_UserNbr = df_tmp['UserNbr'].unique()
for user in all_UserNbr:
df_user = df_tmp.loc[df_tmp['UserNbr'] == user]
if df_user.shape[0] > 0:
open_combined = df_user_open.iloc[0, spcltycol] # capture 1st row
for row in range(1, df_user.shape[0]): # union with any subsequent rows
open_combined = open_combined.union(df_user.iloc[row, spcltycol])
new_row = df_user.drop(['Spclty', 'StartDt'], axis = 1).iloc[0].tolist()
new_row.append(open_combined)
out_rows.append(new_row)
# construct new dataframe with no redundant UserID rows:
df_out = pd.DataFrame(out_rows,
columns = ['UserNbr', 'Spclty'])
# convert Spclty sets to dicts:
df_out['Spclty'] = [dict(df_out['Spclty'][row]) for row in range(df_out.shape[0])]
The conversion to dict gets rid of specialties that are repeated between rows, In the output, a Spclty value should look like this:
{'104': '2010-01-31', '215': '2014-11-21', '352': '2016-07-13'}
except that there may be more key-value pairs than in any corresponding input row (resulting from aggregation over UserNbr).
I withdraw this question.
I had hoped there was an efficient way to use groupby with something else, but I haven't found any examples with a complex data structure like this one and have received no guidance.
For anyone who gets similarly stuck with very slow aggregation problems in Python, I suggest stepping up to PySpark. I am now tackling this problem with a Databricks notebook and am making headway with the pyspark.sql.window Window functions. (Now, it only takes minutes to run a test instead of hours!)
A partial solution is in the answer here:
PySpark list() in withColumn() only works once, then AssertionError: col should be Column

Compare two lists and create multiple lists based on matching data

I have a list of 20 movies, saved like this:
library = [
['name1','director1','genre1',running_time_in_mins_1],
['name2','director2','genre2',running_time_in_mins_2],
etc up to 20
Based on that library, I need to be able to show the average running time for all movies of a particular genre. I would be looking to output something like: "Average running time for xyz is ### minutes". There could be any number of genres (6 currently in my library) or movies in the library.
I can create a list of unique genres using
genre=[]
for y in range (0,len(library)):
if (library[y][2]) not in genre:
genre.append(library[y][2])
I then thought I would create separate lists for each genre and add the matching run time but this doesn't work
for x in range (len(library)):
for z in range(0,(len(genre))):
if library[x][2] == genre[z]:
z=[]
z.append(library[x][3])
print(z)
pandas is a good library that does what you want, though it’s a little tricky to learn.
For now, try:
for genre in set(row[2] for row in library):
times = [row[3] for row in library if row[2]==genre]
print("average runtime for", genre, "is", sum(times)/len(times))
List comprehensions are a very useful tool
Getting a list of unique directors:
>>> library = [['n1', 'd1', 30], ['n2', 'd2', 40], ['n1', 'd2', 20]]
>>> set([x[1] for x in library])
set(['d2', 'd1'])
Getting a list of times for a specific name
>>> times = [x[2] for x in library if x[0] == 'n1']
[30, 20]
Calculating an average time
>>> sum(times) / len(times) # python 3
25
But list of list is really not optimal and makes the code unreadable (what's x[0]?). Depending on how you get the data, what else you want to do with it and how much time you want to spend learning more python/libraries, I would suggest: list of dictionaries, list of class, pandas dataframe or a database.

List, tuples or dictionary, differences and usage, How can I store info in python

I'm very new in python (I usually write in php). I want to understand how to store information in an associative array, and if you can explain me whats the difference of "tuples", "arrays", "dictionary" and "list" will be wonderful (I tried to read different source but I still not caching it).
So This is my code:
#!/usr/bin/python3.4
import csv
import string
nidless_keys = dict()
nidless_keys = ['test_string1','test_string2'] #this contain the string to
# be searched in linesreader
data = {'type':[],'id':[]} #here I want to store my information
with open('path/to/csv/file.csv',newline="") as csvfile:
linesreader = csv.reader(csvfile,delimiter=',',quotechar="|")
for row in linesreader: #every line in this csv have a url like
#www.test.com/?test_string1&id=123456
current_row_string = str(row)
for needle in nidless_keys:
current_needle = str(needle)
if current_needle in current_row_string:
data[current_needle[current_row_string[-8:]]) += 1 # also I
#need to count per every id how much rows there are.
In conclusion:
my_data_stored = [current_needle][current_row_string[-8]]
current_row_string[-8] is a url which the last 8 digit of the url is an ID.
So the array should looks like this at the end of the script:
test_string1 = 123456 = 20
= 256468 = 15
test_string2 = 123155 = 10
Edit 1:
Which type I need here to store the information?
Can you tell me how to resolve this script?
It seems you want to count how many times an ID in combination with a test string occurs.
There can be multiple ID/count combinations associated with every test string.
This suggests that you should use a dictionary indexed by the test strings to store the results. In that dictionary I would suggest to store collections.Counter objects.
This way, you would have to add a special case when a key in the results dictionary isn't found to add an empty Counter. This is a common problem, so there is a specialized form of dictionary in the collections module called defaultdict.
import collections
import csv
# Using a tuple for the keys so it cannot be accidentally modified
keys = ('test_string1', 'test_string2')
result = collections.defaultdict(collections.Counter)
with open('path/to/csv/file.csv',newline="") as csvfile:
linesreader = csv.reader(csvfile,delimiter=',',quotechar="|")
for row in linesreader:
for key in keys:
if key in row:
id = row[-6:] # ID's are six digits in your example.
# The first index is into the dict, the second into the Counter.
result[key][id] += 1
There is an even easier way, by using regular expressions.
Since you seem to treat every row in a CSV file as a string, there is little need to use the CSV reader, so I'll just read the whole file as text.
import re
with open('path/to/csv/file.csv') as datafile:
text = datafile.read()
pattern = r'\?(.*)&id=(\d+)'
The pattern is a regular expression. This is a large topic in and of itself, so I'll only cover briefly what it does. (You might also want to check out the relevant HOWTO) At first glance it looks like complete gibberish, but it is actually a complete language.
In looks for two things in a line. Anything between ? and &id=, and a sequence of digits after &id=.
I'll be using IPython to give an example.
(If you don't know it, check out IPython. It is great for trying things and see if they work.)
In [1]: import re
In [2]: pattern = r'\?(.*)&id=(\d+)'
In [3]: text = """www.test.com/?test_string1&id=123456
....: www.test.com/?test_string1&id=123456
....: www.test.com/?test_string1&id=234567
....: www.test.com/?foo&id=234567
....: www.test.com/?foo&id=123456
....: www.test.com/?foo&id=1234
....: www.test.com/?foo&id=1234
....: www.test.com/?foo&id=1234"""
The text variable points to the string which is a mock-up for the contents of your CSV file.
I am assuming that:
every URL is on its own line
ID's are a sequence of digits.
If these assumptions are wrong, this won't work.
Using findall to extract every match of the pattern from the text.
In [4]: re.findall(pattern, test)
Out[4]:
[('test_string1', '123456'),
('test_string1', '123456'),
('test_string1', '234567'),
('foo', '234567'),
('foo', '123456'),
('foo', '1234'),
('foo', '1234'),
('foo', '1234')]
The findall function returns a list of 2-tuples (that is key, ID pairs). Now we just need to count those.
In [5]: import collections
In [6]: result = collections.defaultdict(collections.Counter)
In [7]: intermediate = re.findall(pattern, test)
Now we fill the result dict from the list of matches that is the intermediate result.
In [8]: for key, id in intermediate:
....: result[key][id] += 1
....:
In [9]: print(result)
defaultdict(<class 'collections.Counter'>, {'foo': Counter({'1234': 3, '123456': 1, '234567': 1}), 'test_string1': Counter({'123456': 2, '234567': 1})})
So the complete code would be:
import collections
import re
with open('path/to/csv/file.csv') as datafile:
text = datafile.read()
result = collections.defaultdict(collections.Counter)
pattern = r'\?(.*)&id=(\d+)'
intermediate = re.findall(pattern, test)
for key, id in intermediate:
result[key][id] += 1
This approach has two advantages.
You don't have to know the keys in advance.
ID's are not limited to six digits.
A brief summary of the python data types you mentioned:
A dictionary is an associative array, aka hashtable.
A list is a sequence of values.
An array is essentially the same as a list, but limited to basic datatypes. My impression is that they only exists for performance reasons, don't think I've ever used one. If performance is that critical to you, you probably don't want to use python in the first place.
A tuple is a fixed-length sequence of values (whereas lists and arrays can grow).
Lets take them one by one.
Lists:
List is a very naive kind of data structure similar to arrays in other languages in terms of the way we write them like:
['a','b','c']
This is a list in python , but seems very similar to array structure.
However there is a very large difference in the way lists are used in python and the usual arrays.
Lists are heterogenous in nature. This means that we can store any kind of data simultaneously inside it like:
ls = [1,2,'a','g',True]
As you can see, we have various kinds of data within a list and is a valid list.
However, one important thing about them is that we can access the list items using zero based indices. So we can write:
print ls[0],ls[3]
output: 1 g
Dictionary:
This datastructure is similar to a hash map data structure. It contains a (key,Value) pair. An empty dictionary looks like:
dc = {}
Now, to store a key,value pair, e.g., ('potato',3),(tomato,5), we can do as:
dc['potato'] = 3
dc['tomato'] = 5
and we saved the data in the dictionary dc.
The important thing is that we can even store another data structure element like a list within a dictionary like:
dc['list1'] = ls , where ls is the list defined above.
This shows the power of using dictionary.
In your case, you have difined a dictionary like this:
data = {'type':[],'id':[]}
This means that your dictionary will consist of only two keys and each key corresponds to a list, which are empty for now.
Talking a bit about your script, the expression :
current_row_string[-8:]
doesn't make a sense. The index should have been -6 instead of -8 that would give you the id part of the current row.
This part is the id and should have been stored in a variable say :
id = current_row_string[-6:]
Further action can be performed as seen the answer given by Roland.

Resources