A note on Get-HtmlTable

[scriptingstudio]

Line 53

$rows =    $h.SelectNodes("//table[$TableIndex]//tr")

This XPath selector says, "Select the $TableIndex table element and all tr elements no matter where they are under the table element".
There are strange web pages, like https://docs.microsoft.com/en-us/windows/release-health/windows11-release-information, where the selector dont work. Weirdness is that multiple tables selected.
A work around for this:

$rows = $htmldoc.SelectSingleNode("//table[$TableIndex]").SelectNodes(".//tr")

Anybody tested the original selector on pages with multiple tables?

[dfinke]

can you do a PR?

[scriptingstudio]

can you do a PR?

Done

Further notes to discuss:

• ambiguity of $TableIndex parameter
  In Windows part it clearly implies a numeric type. For PS7 it can be an XPath expression. I would suggest to make it [int] type or introduce type validation in Windows part

if (-not ($TableIndex -as [int])) { # auto adjust
    Write-Warning "TableIndex must be numeric type. Reset to 0."
    $TableIndex = 0
}

• line 43 and [missing] value. This results in table structure corruption. I would suggest to give it a unique name

if(!$propertyName) { $propertyName= "P$counter"}

• Table header names duplicates - quite a rare case but it results in table structure corruption as well. I would suggest to test names for dups after condition if (-not $propertyNames) {}

$test = $propertyNames | Group-Object -NoElement | Where-Object Count -gt 1
if ($test) {
    Write-Warning "Table header has duplicates: $($test.name -join ', ')."
    return
}

• $propertyNames calculation in Windows part is not optimal considering a dups test. Needs to be taken out of the cycle

$table     = $r.ParsedHtml.getElementsByTagName('table')[$TableIndex]
$totalRows = @($table.rows).count
$cells     = @($table.rows[0].cells)
$headless  = $cells[0].tagName -ne 'th'
if (-not $propertyNames) {
    $propertyNames = if (-not $headless) {
        @($cells.ForEach{$_.innertext -replace ' '})
    } else {
        @((1..($cells.Count + 2)).Foreach{"P$_"})
    }
}
if ($FirstDataRow -eq 0 -and -not $headless) {$FirstDataRow = 1} # auto adjust index

[dfinke]

thanks for the notes