I've been trying to create a dictionary of data frames so I can store data coming from different files. I have one dataframe created in the following loop, and I would like to aggregate them to have each dataframe to the dictionary. I will have to join them later by the date.
d = {}
for num in range(3,14):
nodeName = "rgs" + str(num).zfill(2) #The key should be the nodeName
# Bunch of stuff to get the data ...
# Fill dataframe
data = {'date':date_list, 'users':users_list}
df = pd.DataFrame(data)
df = df.convert_objects(convert_numeric=True)
df = df.dropna(subset=['users'])
df['users'] = df['users'].astype(int)
d = {nodeName:df}
print d
The problem that I have is, if I print the dictionary out of the loop I only have one item, the last one.
{'rgs13': date users
0 2016-01-18 1
1 2016-01-19 1
2 2016-01-20 1
3 2016-01-21 1
4 2016-01-22 1
5 2016-01-23 1
6 2016-01-24 0
But I can clearly see that I can generate all the dataframes without problems inside the loop. How can I make the dictionary to keep all the df's? What am I doing wrong?
Thanks for the help.
It's because in the end you are re-defining d.
What you want is this:
d = {}
for num in range(3,14):
nodeName = "rgs" + str(num).zfill(2) #The key should be the nodeName
# Bunch of stuff to get the data ...
# Fill dataframe
data = {'date':date_list, 'users':users_list}
df = pd.DataFrame(data)
df = df.convert_objects(convert_numeric=True)
df = df.dropna(subset=['users'])
df['users'] = df['users'].astype(int)
d[nodeName] = df
print d
Instead of d = {nodeName:df} use
d[nodeName] = df
Since this adds a key/value pair to d whereas d = {nodeName:df} reassigns d to a new dict (with only the one key/value pair). Doing that in a loop spells death to all the previous key/value pairs.
You may find Ned Batchelder's Facts and myths about Python names and values a useful read. It will give you the right mental model for thinking about the relationship between variable names and values, and help you see what statements modify values (e.g. d[nodeName] = df) versus reassign variable names (e.g. d = {nodeName:df}).
Related
I am getting myself very confused over a problem I am encountering with a short python script I am trying to put together. I am trying to iterate through a dataframe, appending rows to a new dataframe, until a certain value is encountered.
import pandas as pd
#this function will take a raw AGS file (saved as a CSV) and convert to a
#dataframe.
#it will take the AGS CSV and print the top 5 header lines
def AGS_raw(file_loc):
raw_df = pd.read_csv(file_loc)
#print(raw_df.head())
return raw_df
import_df = AGS_raw('test.csv')
def AGS_snip(raw_df):
for i in raw_df.iterrows():
df_new_row = pd.DataFrame(i)
cut_df = pd.DataFrame(raw_df)
if "**PROJ" == True:
cut_df = cut_df.concat([cut_df,df_new_row],ignore_index=True, sort=False)
elif "**ABBR" == True:
break
print(raw_df)
return cut_df
I don't need to get into specifics, but the values (**PROJ and **ABBR) in this data occur as single cells as the top of tables. So I want to loop row-wise through the data, appending rows until **ABBR is encountered.
When I call AGS_snip(import_df), nothing happens. Previous incarnations just spat out the whole df, and I'm just confused over the logic of the loops. Any assistance much appreciated.
EDIT: raw text of the CSV
**PROJ,
1,32
1,76
32,56
,
**ABBR,
1,32
1,76
32,56
The test CSV looks like this:
The reason that "nothing happens" is likely b/c of the conditions you're using in if and elif.
Neither "**PROJ" == True nor "**ABBR" == True will ever be True because neither "**PROJ" nor "**ABBR" are equal to True. Your code is equivalent to:
def AGS_snip(raw_df):
for i in raw_df.iterrows():
df_new_row = pd.DataFrame(i)
cut_df = pd.DataFrame(raw_df)
if False:
cut_df = cut_df.concat([cut_df,df_new_row],ignore_index=True, sort=False)
elif False:
break
print(raw_df)
return cut_df
Which is the same as:
def AGS_snip(raw_df):
for i in raw_df.iterrows():
df_new_row = pd.DataFrame(i)
cut_df = pd.DataFrame(raw_df)
print(raw_df)
return cut_df
You also always return from inside the loop and df_new_row isn't used for anything, so it's equivalent to:
def AGS_snip(raw_df):
first_row = next(raw_df.iterrows(), None)
if first_row:
cut_df = pd.DataFrame(raw_df)
print(raw_df)
return cut_df
Here's how to parse your CSV file into multiple separate dataframes based on a row condition. Each dataframe is stored in a Python dictionary, with titles as keys and dataframes as values.
import pandas as pd
df = pd.read_csv('ags.csv', header=None)
# Drop rows which consist of all NaN (Not a Number) / missing values.
# Reset index order from 0 to the end of dataframe.
df = df.dropna(axis='rows', how='all').reset_index(drop=True)
# Grab indices of rows beginning with "**", and append an "end" index.
idx = df.index[df[0].str.startswith('**')].append(pd.Index([len(df)]))
# Dictionary of { dataframe titles : dataframes }.
dfs = {}
for k in range(len(idx) - 1):
table_name = df.iloc[idx[k],0]
dfs[table_name] = df.iloc[idx[k]+1:idx[k+1]].reset_index(drop=True)
# Print the titles and tables.
for k,v in dfs.items():
print(k)
print(v)
# **PROJ
# 0 1
# 0 1 32.0
# 1 1 76.0
# 2 32 56.0
# **ABBR
# 0 1
# 0 1 32.0
# 1 1 76.0
# 2 32 56.0
# Access each dataframe by indexing the dictionary "dfs", for example:
print(dfs['**ABBR'])
# 0 1
# 0 1 32.0
# 1 1 76.0
# 2 32 56.0
# You can rename column names with for example this code:
dfs['**PROJ'].set_axis(['data1', 'data2'], axis='columns', inplace=True)
print(dfs['**PROJ'])
# data1 data2
# 0 1 32.0
# 1 1 76.0
# 2 32 56.0
I have many different tables that all have different column names and each refer to an outcome, like glucose, insulin, leptin etc (except keep in mind that the tables are all gigantic and messy with tons of other columns in them as well).
I am trying to generate a report that starts empty but then adds columns based on functions applied to each of the glucose, insulin, and leptin tables.
I have included a very simple example - ignore that the function makes little sense. The below code works, but I would like to, instead of copy + pasting final_report["outcome"] = over and over again, just run the find_result function over each of glucose, insulin, and leptin and add the "glucose_result", "insulin_result" and "leptin_result" to the final_report in one or a few lines.
Thanks in advance.
import pandas as pd
ids = [1,1,1,1,1,1,2,2,2,2,2,2,3,3,3,4,4,4,4,4,4]
timepoint = [1,2,3,4,5,6,1,2,3,4,5,6,1,2,4,1,2,3,4,5,6]
outcome = [2,3,4,5,6,7,3,4,1,2,3,4,5,4,5,8,4,5,6,2,3]
glucose = pd.DataFrame({'id':ids,
'timepoint':timepoint,
'outcome':outcome})
insulin = pd.DataFrame({'id':ids,
'timepoint':timepoint,
'outcome':outcome})
leptin = pd.DataFrame({'id':ids,
'timepoint':timepoint,
'outcome':outcome})
ids = [1,2,3,4]
start = [1,1,1,1]
end = [6,6,6,6]
final_report = pd.DataFrame({'id':ids,
'start':start,
'end':end})
def find_result(subject, start, end, df):
df = df.loc[(df["id"] == subject) & (df["timepoint"] >= start) & (df["timepoint"] <= end)].sort_values(by = "timepoint")
return df["timepoint"].nunique()
final_report['glucose_result'] = final_report.apply(lambda x: find_result(x['id'], x['start'], x['end'], glucose), axis=1)
final_report['insulin_result'] = final_report.apply(lambda x: find_result(x['id'], x['start'], x['end'], insulin), axis=1)
final_report['leptin_result'] = final_report.apply(lambda x: find_result(x['id'], x['start'], x['end'], leptin), axis=1)
If you have to use this code structure, you can create a simple dictionary with your dataframes and their names and loop through them, creating new columns with programmatically assigned names:
input_dfs = {"glucose": glucose, "insulin": insulin, "leptin": leptin}
for name, df in input_dfs.items():
final_report[f"{name}_result"] = final_report.apply(
lambda x: find_result(x['id'], x['start'], x['end'], df),
axis=1
)
Output:
id start end glucose_result insulin_result leptin_result
0 1 1 6 6 6 6
1 2 1 6 6 6 6
2 3 1 6 3 3 3
3 4 1 6 6 6 6
I am trying to categorize a dataset based on the string that contains the name of the different objects of the dataset.
The dataset is composed of 3 columns, df['Name'], df['Category'] and df['Sub_Category'], the Category and Sub_Category columns are empty.
For each row I would like to check in different lists of words if the name of the object contains at least one word in one of the list. Based on this first check I would like to attribute a value to the category column. If it finds more than 1 word in 2 different lists I would like to attribute 2 values to the object in the category column.
Moreover, I would like to be able to identify which word has been checked in which list in order to attribute a value to the sub_category column.
Until now, I have been able to do it with only one list, but I am not able to identity which word has been checked and the code is very long to run.
Here is my code (where I added an example of names found in my dataset as df['Name']) :
import pandas as pd
import numpy as np
df['Name'] = ['vitrine murale vintage','commode ancienne', 'lustre antique', 'solex', 'sculpture médievale', 'jante voiture', 'lit et matelas', 'turbine moteur']
furniture_check = ['canape', 'chaise', 'buffet','table','commode','lit']
vehicle_check = ['solex','voiture','moto','scooter']
art_check = ['tableau','scuplture', 'tapisserie']
for idx, row in df.iterrows():
for c in furniture_check:
if c in row['Name']:
df.loc[idx, 'Category'] = 'Meubles'
Any help would be appreciated
Here is an approach that expands lists, merges them and re-combines them.
df = pd.DataFrame({"name":['vitrine murale vintage','commode ancienne', 'lustre antique', 'solex', 'sculpture médievale', 'jante voiture', 'lit et matelas', 'turbine moteur']})
furniture_check = ['canape', 'chaise', 'buffet','table','commode','lit']
vehicle_check = ['solex','voiture','moto','scooter']
art_check = ['tableau','scuplture', 'tapisserie']
# put categories into a dataframe
dfcat = pd.DataFrame([{"category":"furniture","values":furniture_check},
{"category":"vechile","values":vehicle_check},
{"category":"art","values":art_check}])
# turn apace delimited "name" column into a list
dfcatlist = (df.assign(name=df["name"].apply(lambda x: x.split(" ")))
# explode list so it can be used as join. reset_index() to keep a copy of index of original DF
.explode("name").reset_index()
# merge exploded names on both side
.merge(dfcat.explode("values"), left_on="name", right_on="values")
# where there are multiple categoryies, make it a list
.groupby("index", as_index=False).agg({"category":lambda s: list(s)})
# but original index back...
.set_index("index")
)
# simple join and have names and list of associated categories
df.join(dfcatlist)
name
category
0
vitrine murale vintage
nan
1
commode ancienne
['furniture']
2
lustre antique
nan
3
solex
['vechile']
4
sculpture médievale
nan
5
jante voiture
['vechile']
6
lit et matelas
['furniture']
7
turbine moteur
nan
I have an original dataset with informations stored as a list of dict, in a column (this is a mongodb extract). This is the column :
[{u'domain_id': ObjectId('A'), u'p': 1},
{u'domain_id': ObjectId('B'), u'p': 2},
{u'domain_id': ObjectId('B'), u'p': 3},
...
{u'domain_id': ObjectId('CG'), u'p': 101}]
I'm only interested in the first 10 dict ( 'p' value from 1 to 10). The output dataframe should look like this :
index | A | ... | B
------------------------
0 | 1 | ... | 2
1 | Nan | ... | Nan
2 | Nan | ... | 3
e.g : For each line of my original DataFrame, I create a column for each domain_id, and I associate it with the corresponding 'p' value. I can have the same domain_id for several 'p' value, in this case I only keep the first one (smaller 'p')
Here is my current code, which may be easier to understand :
first = True
for i in df.index[:]: # for each line of original Dataframe
temp_list = df["positions"][i] # this is the column with the list of dict inside
col_list = []
data_list = []
for j in range(10): # get the first 10 values
try:
if temp_list[j]["domain_id"] not in col_list: # check if domain_id already exist
col_list.append(temp_list[j]["domain_id"])
data_list.append(temp_list[j]["p"])
except IndexError as e:
print e
df_temp = pd.DataFrame([np.transpose(data_list)],columns = col_list) # create a temporary DataFrame for this line of the original DataFrame
if first:
df_kw = df_temp
first = False
else:
# pass
df_kw = pd.concat([df_kw,df_temp], axis=0, ignore_index=True) # concat all the temporary DataFrame : now I have my output Dataframe, with the same number of lines as my original DataFrame.
This is all working fine, but it is very very slow as I have 15k lines and end up with 10k columns.
I'm sure (or at least I hope very much) that there is a simpler an faster solution : any advice will be much appreciated.
I found a decent solution : the slow part is the concatenation, so it is way more efficient to first create the dataframe and then update the values.
Create the DataFrame:
for i in df.index[:]:
temp_list = df["positions"][i]
for j in range(10):
try:
# if temp_list[j]["domain_id"] not in col_list:
col_list.append(temp_list[j]["domain_id"])
except IndexError as e:
print e
df_total = pd.DataFrame(index=df.index, columns=set(col_list))
Update the values :
for i in df.index[:]:
temp_list = df["positions"][i]
col_list = []
for j in range(10):
try:
if temp_list[j]["domain_id"] not in col_list: # avoid overwriting values
df_total.loc[i, temp_list[j]["domain_id"]] = temp_list[j]["p"]
col_list.append(temp_list[j]["domain_id"])
except IndexError as e:
print e
Creating a 15k x 6k DataFrame took about 6 seconds on my computer, and filling it took 27 seconds.
I killed the former solution after more than 1 hour running, so this is really faster.
I am working with the sklearn.datasets.fetch_20newsgroups() dataset. Here, there are some documents that belong to more than one news group. I want to treat those documents as two different entities that each belong to one news group. To do this, I've brought the document IDs and group names into a dataframe.
import sklearn
from sklearn import datasets
data = datasets.fetch_20newsgroups()
filepaths = data.filenames.astype(str)
keys = []
for path in filepaths:
keys.append(os.path.split(path)[1])
groups = pd.DataFrame(keys, columns = ['Document_ID'])
groups['Group'] = data.target
groups.head()
>> Document_ID Group
0 102994 7
1 51861 4
2 51879 4
3 38242 1
4 60880 14
print (len(groups))
>>11314
print (len(groups['Document_ID'].drop_duplicates()))
>>9840
print (len(groups['Group'].drop_duplicates()))
>>20
For each Document_ID, I want to change its value if it has more than one Group number assigned. Example,
groups[groups['Document_ID']=='76139']
>> Document_ID Group
5392 76139 6
5680 76139 17
I want this to become:
>> Document_ID Group
5392 76139 6
5680 12345 17
Here, 12345 is a random new ID that is not already in keys list.
How can I do this?
You can find all the rows that contain duplicate Document_ID after the first with the duplicated methdod. Then create a list of new id's beginning with one more than the max id. Use the loc indexing operator to overwrite the duplicate keys with the new ids.
groups['Document_ID'] = groups['Document_ID'].astype(int)
dupes = groups.Document_ID.duplicated(keep='first')
max_id = groups.Document_ID.max() + 1
new_id = range(max_id, max_id + dupes.sum())
groups.loc[dupes, 'Document_ID'] = new_id
Test case
groups.loc[[5392,5680]]
Document_ID Group
5392 76139 6
5680 179489 17
Ensure that no duplicates remain.
groups.Document_ID.duplicated(keep='first').any()
False
Kinda Hacky, but why not!
data = {"Document_ID": [102994,51861,51879,38242,60880,76139,76139],
"Group": [7,1,3,4,4,6,17],
}
groups = pd.DataFrame(data)
groupDict ={}
tempLst=[]
#Create a list of unique ID's
DocList = groups['Document_ID'].unique()
DocList.tolist()
#Build a dictionary and push all group ids to the correct doc id
DocDict = {}
for x in DocList:
DocDict[x] = []
for index, row in groups.iterrows():
DocDict[row['Document_ID']].append(row['Group'])
#For all doc Id's with multip entries create a new id with the group id as a decimal point.
groups['DupID'] = groups['Document_ID'].apply(lambda x: len(DocDict[x]))
groups["Document_ID"] = np.where(groups['DupID'] > 1, groups["Document_ID"] + groups["Group"]/10,groups["Document_ID"])
Hope that helps...