Low-level python skills here (learned programming with SAS).
I am trying to apply a series of fuzzy string matching (fuzzywuzzy lib) formulas on pairs of strings, stored in a base dataframe. Now I'm conflicted about the way to go about it.
Should I write a loop that creates a specific dataframe for each formula and then append all these sub-dataframes in a single one? The trouble with this approach seems to be that, since I cannot dynamically name the sub-dataframe, the resulting value gets overwritten at each turn of the loop.
Or should I create one dataframe in a single loop, taking my formulas names and expression as a dict? The trouble here gives me the same problem as above.
Here is my formulas dict:
# ratios dict: all ratios names and functions
ratios = {"ratio": fuzz.ratio,
"partial ratio": fuzz.partial_ratio,
"token sort ratio": fuzz.token_sort_ratio,
"partial token sort ratio": fuzz.partial_token_sort_ratio,
"token set ratio": fuzz.token_set_ratio,
"partial token set ratio": fuzz.partial_token_set_ratio
}
And here is the loop I am currently sweating over:
# for loop iterating over ratios
for r, rn in ratios.items():
# fuzzing function definition
def do_the_fuzz(row):
return rn(row[base_column], row[target_column])
# new base df containing ratio data and calculations for current loop turn
df_out1 = pd.DataFrame(data = df_out, columns = [base_column, target_column, 'mesure', 'valeur', 'drop'])
df_out1['mesure'] = r
df_out1['valeur'] = df_out.apply(do_the_fuzz, axis = 1)
It gives me the same problem, namely that the 'mesure' column gets overwritten, and I end up with a column full of the last value (here: 'partial token set').
My overall problem is that I cannot understand if and how I can dynamically name dataframes, columns or values in a python loop (or if I'm even supposed to do it).
I've been trying to come up with a solution myself for too long and I just can't figure it out. Any insight would be very much appreciated! Many thanks in advance!
I would create a dataframe that is updated at each loop iteration:
final_df = pd.DataFrame()
for r, rn in ratios.items():
...
df_out1 = pd.DataFrame(data = df_out, columns = [base_column, target_column, 'mesure', 'valeur', 'drop'])
df_out1['mesure'] = r
df_out1['valeur'] = df_out.apply(do_the_fuzz, axis = 1)
final_df = pd.concat([final_dfl, df_out1], axis=0)
I hope this can help you.
Related
I'have a Dataframe, that correspond to lat/long of an object in movement.
This object go from one place to another, and I created a column that reference what place he is at every second.
I want to split that dataframe, so when the object go in one place, the leave to another, I'll have two separate dataframe.
'None' mean he is between places
My actual code :
def cut_df2(df):
df_copy = df.copy()
#check if change of place
df_copy['changed'] = df_copy['place'].ne(df_copy['place'].shift().bfill()).astype(int)
last = 0
dfs= []
for num, line in df_copy.iterrows():
if line.changed:
dfs.append(df.iloc[last:num,:])
last = num
# Check if last line was in a place
if line.place != 'None':
dfs.append(df.iloc[last:,:])
df_outs= []
# Delete empty dataframes
for num, dataframe in enumerate(dfs):
if not dataframe.empty :
if dataframe.reset_index().place.iloc[0] != 'None':
df_outs.append(dataframe)
return df_outs
It won't work on big dataset, but work on simple examples and I've no idea why, anyone can help me?
Try using this instead:
https://www.geeksforgeeks.org/split-pandas-dataframe-by-rows/
iloc can be a good way to split a dataframe
df1 = datasX.iloc[:, :72]
df2 = datasX.iloc[:, 72:]
Nested for loop is very time inefficient. I have some ideas to make this efficient. Wondering if better alternatives can be shared.
I am trying to create a dataframe in python pulling values from multiple other dataframes. For a small number of variables/columns I can perform simple assignments. In the example below I want a cell each in two dataframes to be compared and make an assignment if equal. If they are not equal I need to iterate through the second dataframe till every cell is evaluated before making any assignment.
"""iterated through each row of first dataframe and then the second. This is to control for values in compared column
are matched correctly. """
for i in range(len(df10)):
for j in range(len(df6)): # this is not an efficient way to perform this action.
if df10.iloc[i,0] == df6.iloc[j,1]:
df10.iloc[i,23] = df6.iloc[j,6]
df10.iloc[i,24] = df6.iloc[j,1]
df10.sample(n=5)
Here is how you can do it, please see comment for description. Leave comment if something is not clear
np.random.seed(10)
df10 = pd.DataFrame(np.random.choice(5, (5,5)))
df6 = pd.DataFrame(np.random.choice(5, (4,6)))
display(df10)
display(df6)
## compare each pair of rows from 0th column of df10 and 1st column of df6
## using numpy broadcast. Which will return matrix of boolean with true at
## element i,j where values are equal
cond = df10.iloc[:,0].values[:,np.newaxis] == df6.iloc[:,1].values
## get matching index in array when the matrix is flatten
indx = np.arange(cond.size)[cond.ravel()]
## convert flattened index to row and colum index (i,j)
## where i crossponds to row index in df10 and j crossponds to
## row index in df6
i,j = indx//len(df6), indx%len(df6)
## set value using fancy indexing
df10.iloc[i,3] = df6.iloc[j,4].values
df10
(I've edited the first column name in the labels_df for clarity)
I have two DataFrames, train_df and labels_df. train_df has integers that map to attribute names in the labels_df. I would like to look up each number within a given train_df cell and return in the adjacent cell, the corresponding attribute name from the labels_df.
So fore example, the first observation in train_df has attribute_ids of 147, 616 and 813 which map to (in the labels_df) culture::french, tag::dogs, tag::men. And I would like to place those strings inside one cell on the same row as the corresponding integers.
I've tried variations of the function below but fear I am wayyy off:
def my_mapping(df1, df2):
tags = df1['attribute_ids']
for i in tags.iteritems():
df1['new_col'] = df2.iloc[i]
return df1
The data are originally from two csv files:
train.csv
labels.csv
I tried this from #Danny :
sample_train_df['attribute_ids'].apply(lambda x: [sample_labels_df[sample_labels_df['attribute_name'] == i]
['attribute_id_num'] for i in x])
*please note - I am running the above code on samples of each DF due to run times on the original DFs.
which returned:
I hope this is what you are looking for. i am sure there's a much more efficient way using look up.
df['new_col'] = df['attribute_ids'].apply(lambda x: [labels_df[labels_df['attribute_id'] == i]['attribute_name'] for i in x])
This is super ugly and one day, hopefully sooner than later, i'll be able to accomplish this task in an elegant fashion though, until then, this is what got me the result I need.
split train_df['attribute_ids'] into their own cell/column
helper_df = train_df['attribute_ids'].str.split(expand=True)
combine train_df with the helper_df so I have the id column (they are photo id's)
train_df2 = pd.concat([train_df, helper_df], axis=1)
drop the original attribute_ids column
train_df2.drop(columns = 'attribute_ids', inplace=True)
rename the new columns
train_df2.rename(columns = {0:'attr1', 1:'attr2', 2:'attr3', 3:'attr4', 4:'attr5', 5:'attr6',
6:'attr7', 7:'attr8', 8:'attr9', 9:'attr10', 10:'attr11'})
convert the labels_df into a dictionary
def create_file_mapping(df):
mapping = dict()
for i in range(len(df)):
name, tags = df['attribute_id_num'][i], df['attribute_name'][i]
mapping[str(name)] = tags
return mapping
map and replace the tag numbers with their corresponding tag names
train_df3 = train_df2.applymap(lambda s: my_map.get(s) if s in my_map else s)
create a new column of the observations tags in a list of concatenated values
helper1['new_col'] = helper1[helper1.columns[0:10]].apply(lambda x: ','.join(x.astype(str)), axis = 1)
I have some data 33k rows x 57 columns.
In some columns there is a data which I want to translate with dictionary.
I have done translation, but now I want to write back translated data to my data set.
I have problem with saving tuples output from for loop.
I am using tuples for creating good translation. .join and .append is not working in my case. I was trying in many case but without any success.
Looking for any advice.
data = pd.read_csv(filepath, engine="python", sep=";", keep_default_na=False)
for index, row in data.iterrows():
row["translated"] = (tuple(slownik.get(znak) for znak in row["1st_service"]))
I just want to see in print(data["1st_service"] a translated data not the previous one before for loop.
First of all, if your csv doesn't already have a 'translated' column, you'll have to add it:
import numpy as np
data['translated'] = np.nan
The problem is the row object you're trying to write to is only a view of the dataframe, it's not the dataframe itself. Plus you're missing square brackets for your list comprehension, if I'm understanding what you're doing. So change your last line to:
data.loc[index, "translated"] = tuple([slownik.get(znak) for znak in row["1st_service"]])
and you'll get a tuple written into that one cell.
In future, posting the exact error message you're getting is very helpful!
I have manage it, below working code:
data = pd.read_csv(filepath, engine="python", sep=";", keep_default_na=False)
data.columns = []
slownik = dict([ ])
trans = ' '
for index, row in data.iterrows():
trans += str(tuple([slownik.get(znak) for znak in row["1st_service"]]))
data['1st_service'] = trans.split(')(')
data.to_csv("out.csv", index=False)
Can you tell me if it is well done?
Maybe there is an faster way to do it?
I am doing it for 12 columns in one for loop, as shown up.
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