Записать новый датафрейм из предыдущего
Есть df, со столбцами. https://wdho.ru/mqQQ . Как сделать новый датафрейм, в котором строки создаются только по принципу: "Md" не равно предыдущему значению (например "Md"=2, предыдущее =1) — пишем в новую строку new_df = "Date" = 5-Jan-18, "State bef" = 210.35K. И после когда "Md" не будет равно следующему значению (например "Md"=2, следующее =3), допишем в строку еще и "State after" = 269.51K И так со всеми.
На выходе получается:
Можно так попробовать:
тогда res будет:
UPDATE (после уточнений задачи)
Вы можете группировать именно так, как вам нужно следующим образом:
Ну элементарно же, если мыслить векторно (что часто требуется при работе с Numpy и Pandas ):
Берём разницу между значениями колонки Md и отбираем в df только строки, где эта разница есть (т.е. не 0). Ну и ещё из-за того, что в первой строке разница будет NA , надо заполнить это NA значением, например 1, перед проверкой.
Update: Не заметил сначала условие про "State after" , подумаю ещё.
Ну тогда такой вариант, только остаётся ещё заполнить последнее значение в колонке "State after" , сейчас там NaN , пока не придумал хороший красивый вариант:
pandas.DataFrame.copy#
When deep=True (default), a new object will be created with a copy of the calling object’s data and indices. Modifications to the data or indices of the copy will not be reflected in the original object (see notes below).
When deep=False , a new object will be created without copying the calling object’s data or index (only references to the data and index are copied). Any changes to the data of the original will be reflected in the shallow copy (and vice versa).
Parameters deep bool, default True
Make a deep copy, including a copy of the data and the indices. With deep=False neither the indices nor the data are copied.
Returns copy Series or DataFrame
Object type matches caller.
When deep=True , data is copied but actual Python objects will not be copied recursively, only the reference to the object. This is in contrast to copy.deepcopy in the Standard Library, which recursively copies object data (see examples below).
While Index objects are copied when deep=True , the underlying numpy array is not copied for performance reasons. Since Index is immutable, the underlying data can be safely shared and a copy is not needed.
Since pandas is not thread safe, see the gotchas when copying in a threading environment.
How to create a new dataframe using the another dataframe
I have created and worked on a DataFrame for a project. It looks like the following:
Critics Items Ratings
My DataFrame has 1M+ rows and 8 columns.
I want to create a new DataFrame where the rows are the unique critics, the columns are the unique items, and the individual cells are the rating a critic has given for the particular item. If the critic has not reviewed the item then I want to add an NA over there.
I tried doing the following for the rows:
For the columns, I saw a lot of answers involving people using
But this doesn’t help me since I can’t list down all the unique item names.
Edit: I fixed the issues by using pivot tables. I am new to pandas and was not aware of something like this existing. The exact syntax I used was
Create a New DataFrame From an Existing DataFrame in Pandas?
Sometimes, we need to copy the existing DataFrame with data and indices. However, copying the whole DataFrame is also another way for there to be a direct relationship created between the old DataFrame and the new DataFrame. If we make any changes in the old DataFrame, it will also affect the new DataFrame or vice-versa.
In this article, we are going to see pandas.DataFrame.copy () method, which is used for copy () dataframe.
If we want to create a new DataFrame from an existing DataFrame, then we can use the copy()method. So, in this article, we are going to see how we can use the Pandas DataFrame.copy() method to create another DataFrame from an existing DataFrame.
The Syntax Is Given Below:
In the syntax above, we can see that there is deep either false and true.
These two values are very important to use the copy() method. Let’s see in the details about these two values.
Deep (True): Whenever we use the copy () method, the deep is true by default. This true value indicates that we have to copy all the data and indices from the existing DataFrame and create a new object. Suppose we do any manipulation to the new DataFrame, it will not affect the old DataFrame or vice-versa, which means there will be no relationship connection between the old and new DataFrame, and both can work independently.
Deep (False): When we keep the value of the deep false, then the copy () creates a new object without the data and index. It will create a reference to the data and index of the original DataFrame. If any manipulation to the original DataFrame, it will also affect the shallow copy DataFrame or vice-versa.
Example 1:
Copy the DataFrame using the deep=True:
Line 2: We import the library Pandas as pd. Here, pd means we are importing the Pandas library with the new namespace name called pd. We can use the pd instead of using the pandas full name.
Line 3 to 10: We created a dict with some keys and values, wherein the values are in the list. After creating the dictionary, we convert that dict to a DataFrame (df) using the DataFrame.from_dict () method.
Line 11 to 12: We are printing our dataframe (df), which shows in the output below.
Line 14: We are creating a copy of the df (DataFrame) from the existing df (DataFrame). Here, we are not using any deep=True because that is by default. And, as shown in deep=True, it will create a new object with all data and indices of the existing DataFrame, and there will be no direct relationship between the copy DataFrame and the old DataFrame.
Line 15 to 16: We are printing our copied DataFrame (dfCopy), and the output is shown below:
Output:
Original DataFrame
TV_Show_name TV_Streaming_name show_Season Main Actor
0 The Walking Dead Netflix 4 Rick Grimes
1 Merlin Fx 10 Mordred
2 little evil Disney Plus 4 Karl C. Miller
3 Money Heist Amazon Prime 5 Sergio Marquina
_________________________________________________________
Copied DataFrame
TV_Show_name TV_Streaming_name show_Season Main Actor
0 The Walking Dead Netflix 4 Rick Grimes
1 Merlin Fx 10 Mordred
2 little evil Disney Plus 4 Karl C. Miller
3 Money Heist Amazon Prime 5 Sergio Marquina
Process finished with exit code 0
Example 2:
In this example, we are going to manipulate the old DataFrame and check whether it will affect the dfCopy DataFrame or not. Here, we are using the deep=True to copy the DataFrame:
# python example_2.py
import pandas as pd
data = { ‘TV_Show_name’ : [ ‘The Walking Dead’ , ‘Merlin’ , ‘little evil’ ,
‘Money Heist’ ] ,
‘TV_Streaming_name’ : [ ‘Netflix’ , ‘Fx’ , ‘Disney Plus’ ,
‘Amazon Prime’ ] ,
‘show_Season’ : [ 4 , 10 , 4 , 5 ] ,
‘Main Actor’ : [ ‘Rick Grimes’ , ‘Mordred’ , ‘Karl C. Miller’ ,
‘Sergio Marquina’ ] }
df = pd.DataFrame.from_dict ( data )
print ( ‘Original DataFrame’ )
print ( df )
print ( ‘_________________________________________________________’ )
dfCopy = df.copy ( )
print ( ‘Copied DataFrame’ )
print ( dfCopy )
print ( ‘_________________________________________________________’ )
print ( "************Manipulation done in the original df***************" )
# Now, we are doing data manipulation in the original dataframe
# we are changing the column (‘TV_Show_name’) values to A,B,C,D
# now, we will see this will affect to the dfCopy dataframe or not
df [ ‘TV_Show_name’ ] = df [ ‘TV_Show_name’ ] .replace ( [ ‘The Walking Dead’ ,
‘Merlin’ , ‘little evil’ , ‘Money Heist’ ] , [ ‘A’ , ‘B’ , ‘C’ , ‘D’ ] )
#Now printing both dfCopy(deep=True) and df (original) dataframe
print ( ‘Original DataFrame’ )
print ( df )
print ( ‘Copied DataFrame’ )
print ( dfCopy )
Line 1 to 18: Explanations are already given in the previous program in Example 1.
Line 23: We replace the original df (dataframe) column ([‘TV_Show_name’]) values into [‘A’,’B’,’C’,’D’]. Now, we will check if this manipulation in the original df (dataframe) will affect the dfCopy (deep=True) or not. As we know already, there is no direct relationship between when we use the deep=True.
Line 27 to 30: We print the original df and copy (dataframe) as shown in the output below. From the output, we can confirm that the changes done in the original DataFrame (df) have no effect on the copy (DataFrame):
Output:
From the above example 2, we can confirm that deep=True value when set, the newly created DataFrame from the existing DataFrame has no direct relationship and can perform manipulation without affecting each other.
Example 3:
In this example, we are going to manipulate the old DataFrame and check whether it will affect the dfCopy DataFrame or not. Here, we are using the deep=False to copy the DataFrame:
# python example_3.py
import pandas as pd
data = { ‘TV_Show_name’ : [ ‘The Walking Dead’ , ‘Merlin’ , ‘little evil’ ,
‘Money Heist’ ] ,
‘TV_Streaming_name’ : [ ‘Netflix’ , ‘Fx’ , ‘Disney Plus’ ,
‘Amazon Prime’ ] ,
‘show_Season’ : [ 4 , 10 , 4 , 5 ] ,
‘Main Actor’ : [ ‘Rick Grimes’ , ‘Mordred’ , ‘Karl C. Miller’ ,
‘Sergio Marquina’ ] }
df = pd.DataFrame.from_dict ( data )
print ( ‘Original DataFrame’ )
print ( df )
print ( ‘_________________________________________________________’ )
dfCopy = df.copy ( deep =False )
print ( ‘Copied DataFrame’ )
print ( dfCopy )
print ( ‘_________________________________________________________’ )
# Now, we are doing data manipulation in the original dataframe
# we are changing the column (‘TV_Show_name’) values to A,B,C,D
# now, we will see this will affect to the dfCopy dataframe or not
df [ ‘TV_Show_name’ ] = df [ ‘TV_Show_name’ ] .replace ( [ ‘The Walking Dead’ ,
‘Merlin’ , ‘little evil’ , ‘Money Heist’ ] , [ ‘A’ , ‘B’ , ‘C’ , ‘D’ ] )
#Now printing both dfCopy(deep=False) and df (original) dataframe
print ( ‘_________________________________________________________’ )
print ( ‘Copied DataFrame’ )
print ( dfCopy )
print ( ‘Original DataFrame’ )
print ( df )
Line 1 to 18: Explanations are already given in the program of Example 1. The one change was done at line no. 15. Now, we are using the deep=False instead deep=True.
Line 23: We replace the original df (DataFrame) column ([‘TV_Show_name’]) values into [‘A’,’B’,’C’,’D’]. Now, we will check if this manipulation in the original df (dataframe) will affect the dfCopy (deep=False) or not. As we know already, there is a direct relationship between when we use the deep=False.
Line 27 to 30: We print the original df and copy (DataFrame) as shown in the output below. From the output, we can confirm that the changes done in the original DataFrame (df) have an effect on the copy (DataFrame). The values of the column ([‘TV_Show_name’]) also change in the copy DataFrame.
Output:
Example_4:
Copy the existing DataFrame using assignment operator, which has same direct relationship issue like deep=False:
# python example_4.py
import pandas as pd
data = { ‘TV_Show_name’ : [ ‘The Walking Dead’ , ‘Merlin’ , ‘little evil’ ,
‘Money Heist’ ] ,
‘TV_Streaming_name’ : [ ‘Netflix’ , ‘Fx’ , ‘Disney Plus’ ,
‘Amazon Prime’ ] ,
‘show_Season’ : [ 4 , 10 , 4 , 5 ] ,
‘Main Actor’ : [ ‘Rick Grimes’ , ‘Mordred’ , ‘Karl C. Miller’ ,
‘Sergio Marquina’ ] }
df = pd.DataFrame.from_dict ( data )
print ( ‘Original DataFrame’ )
print ( df )
print ( ‘_________________________________________________________’ )
dfCopy = df
print ( ‘Copied DataFrame’ )
print ( dfCopy )
print ( ‘_________________________________________________________’ )
# Now, we are doing data manipulation in the original dataframe
# we are changing the column (‘TV_Show_name’) values to A,B,C,D
# now, we will see this will affect to the dfCopy dataframe or not
df [ ‘TV_Show_name’ ] = df [ ‘TV_Show_name’ ] .replace ( [ ‘The Walking Dead’ ,
‘Merlin’ , ‘little evil’ , ‘Money Heist’ ] , [ ‘A’ , ‘B’ , ‘C’ , ‘D’ ] )
#Now printing both dfCopy and df (original) dataframe
print ( ‘_________________________________________________________’ )
print ( ‘Copied DataFrame’ )
print ( dfCopy )
print ( ‘Original DataFrame’ )
print ( df )
Line 15: In the above program Example 4, we direct the Dataframe to another variable without using the copy () method. But this also creates a direct relationship between the original DataFrame and the copied DataFrame like the deep=False. The following output shows that if we change anything in the original DataFrame, then it will also affect the copied DataFrame or vice-versa:
Output:
Conclusion:
In this article, we have seen the correct way to copy the existing DataFrame, and doing this will create a new object with data and indices. As we have seen, when we keep the deep value False, it will create a reference to the data and indices to the new copy DataFrame. So, copy using the assignment operator also works in the same way (deep=False), as we have already seen in this article with the help of an example.
Sometimes we need only some of the columns to copy from the existing DataFrame, not the whole. Then we can use the following method, which is similar to the copy (deep=True) but with the name of the columns:
Be careful. If you have only one column, then you must use double square brackets. Otherwise, it will create a series, not a DataFrame.