String Manipulation: Splitting and Joining Strings

I think it's time to get back to some more basic things, so let's get to it.

The split()method of a string object lets you define a delimiter and "split" the string up into multiple pieces, which are returned as an array.

Let's say I do this:

PS C:\Users\tojo2000> $string = "one two three four"

PS C:\Users\tojo2000> $string.split(" ")

one

two

three

four

As it turns out, though, split() splits on a single whitespace character by default.

Observe:

PS C:\Users\tojo2000> $string = "one two three four six"

PS C:\Users\tojo2000> $string.split()

one

two

three

four

six

What's the practical use for this? Let's say you have a tab-separated values file called people.txt that looks like this:

Doe John California

Doe Jane Texas

Neuman Alfred Nebraska

And you want the output to be First Last, Location. You could do it like so:

PS C:\Users\tojo2000> Get-Content people.txt |

>> %{$data = $_.split("`t"); Write-Output "$($data[1]) $($data[0]), $($data[2])"}

>>

John Doe, California

Jane Doe, Texas

Alfred Neuman, Nebraska

But what if you have a file like this called people2.txt (let's say the first part is some kind of ID)?

2323:Doe John California

827:Doe Jane Texas

982982:Neuman Alfred Nebraska

Now you have two delimiters, but the split() method of a string object takes a string as an argument, so you would have to split twice in order to get each part of the string split into an array. If only there was some way to split on a regular expression... Oh, wait, there is.

PS C:\Users\tojo2000> Get-Content people2.txt |

>> %{$data = [regex]::split($_, '\t|:'); Write-Output "$($data[2]) $($data[1]), $($data[3])"}

>>

John Doe, California

Jane Doe, Texas

Alfred Neuman, Nebraska

('\t|:' is a regular expression for a tab character or a colon)

What if you have an array, and you want to make a string out of it? Then you can join the strings together using join(). Now there are two ways to join a list of strings, depending on whether you're using PowerShell v1 or v2.

The first way works in both v1 and v2:

PS C:\Users\tojo2000> $buncha_strings = 'one', 'two', 'three', 'four', 'five', 'six'

PS C:\Users\tojo2000> [string]::join(' and a ', $buncha_strings)

one and a two and a three and a four and a five and a six

So what practical use would there be? For a completely contrived example, maybe you want to create a new version of people2.txt called people3.txt that uses pipe characters to separate the fields. We'll split the strings up into an array just like before, but then we'll join them back together the way we want them.

PS C:\Users\tojo2000> Get-Content people2.txt |

>> %{$data = [regex]::split($_, '\t|:'); [string]::join('|', $data)} > people3.txt

PS C:\Users\tjo2000> Get-Content .\people3.txt

2323|Doe|John|California

827|Doe|Jane|Texas

982982|Neuman|Alfred|Nebraska

There is one more way that you can join a list of strings, but this method only works in v2: the -join operator. It works the same way as the other join(), but the syntax is slightly different.

For example, here is how you would join a list of strings by tab characters:

PS C:\Users\tojo2000> $list_of_strings -join "`t"

So to wrap this all up, let me give a real-world example. I had a SQL query that I wanted to run that would retrieve the SCCM information for a specific list of computers. I needed to update the list, and I had the computers in a file called machines.txt, with one name per line.

Here's how I slurped up the file to create the new query string:

$query = @"

SELECT * FROM v_R_System

WHERE Netbios_Name0

IN ('{0}')

"@ -f ((Get-Content machines.txt) -join "', '")