I have a large data file and I need to delete rows that end in certain letters.
我有一个大的数据文件,我需要删除在某些字母中结束的行。
Here is an example of the file I'm using:
下面是我正在使用的文件的一个例子:
User Name DN
MB212DA CN=MB212DA,CN=Users,DC=prod,DC=trovp,DC=net
MB423DA CN=MB423DA,OU=Generic Mailbox,DC=prod,DC=trovp,DC=net
MB424PL CN=MB424PL,CN=Users,DC=prod,DC=trovp,DC=net
MBDA423 CN=MBDA423,OU=DNA,DC=prod,DC=trovp,DC=net
MB2ADA4 CN=MB2ADA4,OU=DNA,DC=prod,DC=trovp,DC=netenter code here
Code I am using:
我用代码:
from pandas import DataFrame, read_csv
import pandas as pd
f = pd.read_csv('test1.csv', sep=',',encoding='latin1')
df = f.loc[~(~pd.isnull(f['User Name']) & f['UserName'].str.contains("DA|PL",))]
How do I use regular expression syntax to delete the words that end in "DA" and "PL" but make sure I do not delete the other rows because they contain "DA" or "PL" inside of them?
如何使用正则表达式语法删除在“DA”和“PL”中结束的单词,但确保不删除其他行,因为它们包含“DA”或“PL”?
It should delete the rows and I end up with a file like this:
它应该删除这些行,我最后得到的文件是这样的:
User Name DN
MBDA423 CN=MBDA423,OU=DNA,DC=prod,DC=trovp,DC=net
MB2ADA4 CN=MB2ADA4,OU=DNA,DC=prod,DC=trovp,DC=net
First 3 rows are deleted because they ended in DA and PL.
前3行被删除,因为它们在DA和PL中结束。
3 个解决方案
#1
8
You could use this expression
你可以用这个表达式
df = df[~df['User Name'].str.contains('(?:DA|PL)$')]
It will return all rows that don't end in either DA or PL.
它将返回未在DA或PL中结束的所有行。
The ?:
is so that the brackets would not capture anything. Otherwise, you'd see pandas returning the following (harmless) warning:
是为了使括号不包含任何内容。否则,你会看到大熊猫返回以下(无害的)警告:
UserWarning: This pattern has match groups. To actually get the groups, use str.extract.
Alternatively, using endswith()
and without regular expressions, the same filtering could be achieved by using the following expression:
或者,使用endswith()并且不使用正则表达式,可以使用以下表达式来实现相同的过滤:
df = df[~df['User Name'].str.endswith(('DA', 'PL'))]
As expected, the version without regular expression will be faster. A simple test, consisting of big_df
, which consists of 10001 copies of your original df
:
如预期的那样,没有正则表达式的版本将会更快。一个简单的测试,由big_df组成,其中包含您的原始df的10001个副本:
# Create a larger DF to get better timing results
big_df = df.copy()
for i in range(10000):
big_df = big_df.append(df)
print(big_df.shape)
>> (50005, 2)
# Without regular expressions
%%timeit
big_df[~big_df['User Name'].str.endswith(('DA', 'PL'))]
>> 10 loops, best of 3: 22.3 ms per loop
# With regular expressions
%%timeit
big_df[~big_df['User Name'].str.contains('(?:DA|PL)$')]
>> 10 loops, best of 3: 61.8 ms per loop
#2
2
You can use a boolean mask whereby you check if the last two characters of User_Name
are in not (~
) in a set of two character endings:
您可以使用一个boolean掩码来检查User_Name的最后两个字符是否在两个字符结尾的集合中(~):
>>> df[~df.User_Name.str[-2:].isin(['DA', 'PA'])]
User_Name DN
2 MB424PL CN=MB424PL, CN=Users, DC=prod, DC=trovp, DC=net
3 MBDA423 CN=MBDA423, OU=DNA, DC=prod, DC=trovp, DC=net
4 MB2ADA4 CN=MB2ADA4, OU=DNA, DC=prod, DC=trovp, DC=nete...
#3
0
Instead of regular expressions
, you can use the endswith()
method to check if a string ends with a specific pattern.
您可以使用endswith()方法来检查字符串是否以特定的模式结束,而不是正则表达式。
I.e.:
例如:
for row in rows:
if row.endswith('DA') or row.endswith('PL'):
#doSomething
You should create another df using the filtered data, and then use pd.to_csv()
to save a clean version of your file.
您应该使用过滤后的数据创建另一个df,然后使用pd.to_csv()保存文件的干净版本。
#1
8
You could use this expression
你可以用这个表达式
df = df[~df['User Name'].str.contains('(?:DA|PL)$')]
It will return all rows that don't end in either DA or PL.
它将返回未在DA或PL中结束的所有行。
The ?:
is so that the brackets would not capture anything. Otherwise, you'd see pandas returning the following (harmless) warning:
是为了使括号不包含任何内容。否则,你会看到大熊猫返回以下(无害的)警告:
UserWarning: This pattern has match groups. To actually get the groups, use str.extract.
Alternatively, using endswith()
and without regular expressions, the same filtering could be achieved by using the following expression:
或者,使用endswith()并且不使用正则表达式,可以使用以下表达式来实现相同的过滤:
df = df[~df['User Name'].str.endswith(('DA', 'PL'))]
As expected, the version without regular expression will be faster. A simple test, consisting of big_df
, which consists of 10001 copies of your original df
:
如预期的那样,没有正则表达式的版本将会更快。一个简单的测试,由big_df组成,其中包含您的原始df的10001个副本:
# Create a larger DF to get better timing results
big_df = df.copy()
for i in range(10000):
big_df = big_df.append(df)
print(big_df.shape)
>> (50005, 2)
# Without regular expressions
%%timeit
big_df[~big_df['User Name'].str.endswith(('DA', 'PL'))]
>> 10 loops, best of 3: 22.3 ms per loop
# With regular expressions
%%timeit
big_df[~big_df['User Name'].str.contains('(?:DA|PL)$')]
>> 10 loops, best of 3: 61.8 ms per loop
#2
2
You can use a boolean mask whereby you check if the last two characters of User_Name
are in not (~
) in a set of two character endings:
您可以使用一个boolean掩码来检查User_Name的最后两个字符是否在两个字符结尾的集合中(~):
>>> df[~df.User_Name.str[-2:].isin(['DA', 'PA'])]
User_Name DN
2 MB424PL CN=MB424PL, CN=Users, DC=prod, DC=trovp, DC=net
3 MBDA423 CN=MBDA423, OU=DNA, DC=prod, DC=trovp, DC=net
4 MB2ADA4 CN=MB2ADA4, OU=DNA, DC=prod, DC=trovp, DC=nete...
#3
0
Instead of regular expressions
, you can use the endswith()
method to check if a string ends with a specific pattern.
您可以使用endswith()方法来检查字符串是否以特定的模式结束,而不是正则表达式。
I.e.:
例如:
for row in rows:
if row.endswith('DA') or row.endswith('PL'):
#doSomething
You should create another df using the filtered data, and then use pd.to_csv()
to save a clean version of your file.
您应该使用过滤后的数据创建另一个df,然后使用pd.to_csv()保存文件的干净版本。