从字符串列表中提取子字符串列表

Question

I have a list of product names that have a consistent structure.

product_names = ['brand_name1 product1', 'brand_name2 product2', 'brand_name1 product3', 'brand_name3 product4']

Also, I scraped a list of brands from the site filter:

brand_names = ['brand_name1','brand_name2','brand_name3']

Each element in [brand_names] can be found in several elements in [product_names] because several products can belong to the same brand.

Output:

I would like to extract brand_names from product_names and get a.csv file with two columns: Brand, Product.

Solution:

Thanks everyone, I tried to use list comprehension myself. but coded it totally wrong.

import pandas as pd

product_names = ['brand_name1 product1', 'brand_name2 product2','brand_name1 product3']
brand_names = ['brand_name1','brand_name2','brand_name3']

brands = [i for j in product_names for i in brand_names if i in j]

result = pd.DataFrame(
    {'Brand': brands,
     'Product': product_names
     })

result.to_csv('result.csv', index=False)

Answer 1

You can just try this:

for brand in brand_names:
    for product in product_names:
        if (brand in product):
            print(brand,product)

Or an other solution is to use generator:

matching=[]
for brand in brand_names:
    matching.append([product for product in product_names if brand in product])

Answer 2

Hope this is what you are looking for:

data = []
for i in product_names:
    data.append(i.split())
df = pd.DataFrame(data, columns=["Brand", "Product"])
df.to_csv(csv_file_name)

Output:

    Brand       Product
0   brand_name1 product1
1   brand_name2 product2
2   brand_name1 product3
3   brand_name3 product4

Answer 3

In most general case where there is not pattern in your product name, use this:

product_brand = [i for j in product_names for i in brand_names if i in j]

But if there is a pattern, you should leverage that to speed up the process.

output:

product_brand
['brand_name1', 'brand_name2', 'brand_name1', 'brand_name3']

And to write as columns into a csv file, use this:

import csv

rows = zip(product_names,product_brand)
with open('file.txt', "w") as f:
    writer = csv.writer(f)
    for row in rows:
        writer.writerow(row)

output:

brand_name1 product1,brand_name1
brand_name2 product2,brand_name2
brand_name1 product3,brand_name1
brand_name3 product4,brand_name3

Answer 4

Here's an approach:

Assuming your brand names is a list of unique brands you can try this:

import pandas as pd

# Brand and Product lists
product_names = ['brand_name1 product1', 'brand_name2 product2', 'brand_name1 product3', 'brand_name3 product4']
brand_names = ['brand_name1','brand_name2','brand_name3']

# Empty list to save the results
res_ls = []

# Iterate over each brand
for b in brand_names:
    # Select products for your current brand
    brand_prducts = [i for i in product_names if b in i]
    res_ls.append({
        'brand': b,
        'products': ', '.join(brand_prducts)
    })

# Get result as a pandas dataframe
res_df = pd.DataFrame(res_ls)

# Save your dataframe to csv
res_df.to_csv('/path/to/save', index=False)

This is what the pandas dataframe will look like:

Answer 5

This is my little version. Makes a bit of sorting first and assumes the possibility of white spaces in brand and product name.

Sorting makes things easier an nicer. Use of strip() to avoid problems due to white spaces. However, if the product name has white spaces and by accident some are doubled, this is considered a different brand name. To handle this one might need regular expressions.

product_names = ['brand_name1   product1', 'brand name2 product2', 'brand_name1 product 3', ' brand_name3 product4', 'brand name2 product 2']
prbrand_names = ['brand_name1','brand name2','brand_name3']

product_names = sorted( [ s.strip() for s in product_names ] )
prbrand_names = sorted( [ s.strip() for s in prbrand_names ])

with open( "out.csv", "wb") as fpntr:
    cnt = 0
    for bn in prbrand_names:
        # second case is not tested if first is already false -> no IndexError
        while cnt < len( product_names ) and product_names[cnt].startswith( bn ):
            
            pn = product_names[cnt][len( bn ) : ]
            # pn might have unnecessary spaces that can be stripped
            fpntr.write( "{}, {}\n".format( bn, pn.stip() ) )
            cnt += 1

out.csv is:

brand name2, product 2
brand name2, product2
brand_name1, product1
brand_name1, product 3
brand_name3, product4