如何根据过滤器选择特定的子串长度

Question

I have multiple CSV files with different names, containing today's date, customer number, and then the extension. For example:

2019-01-23 XYZF-105.csv
2019-01-23 ABCD-205.csv
2019-01-23 Different nonstandard name.csv
2019-01-23 ##ABCD-305(Trial).csv

I would like to get the part of the name where it contains the customer number only, like ABCD-305.

Tried using a substring to select 8 characters right from the dot, but that doesn't work for those that have suffix like (Trial). Neither it work 11 characters from the beginning, as it will include the ##. Also, it has to avoid the nonstandard names.

I used

$allitems = Get-ChildItem -Path 'C:\Downloads\Customers\*.csv'
$res = @()
foreach ($item in $allitems){
    $item = $item.Name.substring($item.Name.Length - 12,8)
    $res += $Item
}

This way, for the proper names I get good results, but only if the name of the CSV is like 2019-01-23 ABCD-205.csv.

What should be the way to skip the date, skip the .csv extension and get only results with 8 characters, that have a dash after the 4th character? Thanks in advance

Answer 1

Try the following (PSv3+ syntax):

$res = (Get-ChildItem -Path C:\Downloads\Customers\*.csv).Name | 
         Select-String -CaseSensitive '\b[A-Z]{4}-\d{3}\b' |
           ForEach-Object { $_.Matches[0].Value }

• (Get-ChildItem -Path C:\\Downloads\\Customers\\*.csv).Name outputs the file names of all CSV files in dir. C:\\Downloads\\Customers

• Select-String -CaseSensitive '\\b[AZ]{4}-\\d{3}\\b' uses case-sensitive regex (regular-expression) matching to only select file names that contain 4 ( {4} ) uppercase chars. [AZ] , followed by - , followed by 3 digits ( \\d ), on word boundaries ( \\b )

• The ForEach-Object script block then outputs the part of each matching file name that matched the regex ( $_.Matches[0].Value ), so that only the relevant portions of matching file names are collected in $res , as an array.

Answer 2

This would be a good time to use regex. See https://regex101.com/r/AH00n6/1

and understand the following regex:

.*\s[#]*([A-Z]{4}-[0-9]{3}).*.csv

This is a little extra to capture just the names, but gives more insight into how to control the regex.