Aesthetically Pleasing String Replace Statement - powershell

Question

I currently have code that looks like this:
$onestr = $twostr -replace "one","two" -replace "uno","dos"
I would like to format it like this:
$onestr = $twostr -replace "one","two"
-replace "uno","dos"
such that the replace statements stack on top of each other.
I could use backtic as the line continuation character, but other stackoverflow questions cover why that is not a good idea.
I tried code that looks like this:
$onestr = ($twostr -replace "one","two"
-replace "uno","dos"
)
But I got an error that the paren is not matched.
My actual code has several replace statements (not just two).

Answer 1

If you have a lot of replacements, what about a different approach which allows much more scope for the aligning of the replacement pairs nicely, without affecting the block of replace code.
$onestr = 'one thing uno thing'
$Pairs = #{
'one' = 'two'
'uno' = 'dos'
}
$Pairs.GetEnumerator() | ForEach-Object {
$onestr = $onestr -replace $_.Name, $_.Value
}
$onestr

Answer 2

This is less common, maybe awful, but more exactly the layout you ask about - using the .Net framework string replace method instead of the PowerShell operator.
$onestr = 'one thing uno thing'
$onestr.
Replace('one', 'two').
Replace('uno', 'dos')

Answer 3

If you don't want to use a backtick ` as a line continuation character, then I suggest you just make these separate statements:
$onestr = $twostr -replace "one","two"
$onestr = $onestr -replace "uno","dos"

Answer 4

An athletically, not aesthetically, pleasing method, more obvious than a backtick, might be a comment:
$onestr = $twostr -replace "one","two" <#
#> -replace "uno","dos" <#
#> -replace "foo","bar"
At least it's DRYer than repeating the assignments!
Or the heavy artillery: .NET Regex.Replace with a scriptblock so that the text is iterated only once:
function translate(
[Parameter(Mandatory=$true, ValueFromPipeline=$true)]
[string]$text,
[Parameter(Mandatory=$true)]
[hashtable]$map,
[switch]$caseSensitive
) {
$options = if ([bool]$caseSensitive) { '' } else { '(?i)' }
([regex]($options +
($map.Keys -replace '[$^*?.+|()\[\]{}\\]', '\$&' -join '|')
)).Replace($text, { param([string]$match) $map[$match] })
}
'Abc' | translate -map #{
'a'='1'
'b'='2'
'c'='3'
}
It's case-INsensitive by default.

Answer 5

You might consider using a metafunction with a [scriptblock] parameter to unwrap your expression into one line without any backticks. For example, a simple Invoke-Unwrap metafunction could be used to define ConvertTo-Spanish as follows:*
function Invoke-Unwrap( [scriptblock]$sb) {
Invoke-Expression ($sb -replace '\r?\n\s*-', ' -')
}
Invoke-Unwrap {
function script:ConvertTo-Spanish( [string]$str ) {
$str
-replace "one","uno"
-replace "two","dos"
-replace "thing","lo"
}
}
ConvertTo-Spanish "thing one and thing two"
As written Invoke-Unwrap is very general, it allows word-wrapping before any operator or switch starting with a "-". It could be expanded to allow other operator-characters at the beginning of a line, permitting arithmetic expressions to be written in a word-wrapped form. I'm thinking ($sb -replace '\r?\n\s*([-+*/])', ' $1').
Or, if you want, you can enhance the processing of the string passed to Invoke-Expression to make your own little domain-specific language, perhaps implementing something more SED like.
Apologies for my limited knowledge of Spanish... :-)