```powershell
$rows = $h.SelectNodes("//table[$TableIndex]//tr")
```
XPath selector in line 53 uses complex expression that can lead to unexpected result. The problem is that HtmlAgilityPack may have specific issues. In particular, on websites containing multiple tables this selector can find not one table. This is aggravated by the fact that tables can have different structures.
To avoid ambiguity this PR suggests to separate queries. Oneliner simplifies error checking
```powershell
$rows = try {
$h.SelectSingleNode("//table[$TableIndex]").SelectNodes(".//tr")
} catch {}
if (-not $rows) {Write-Warning "Could not find rows for `"//table[$TableIndex]`" in $Url ."}
```
This expression doesn't even need testing, it just works.
$StartRow parameter had ValidateRange(1,9999) which was limiting the start row to a maximum of 9999. There's no good reason this cannot be any row in the spreadsheet, of which the maximum is 1048576 in modern instances of Excel.
If data in the Excel file contains single quotes, then the insert will bomb as it will be malformed. Quotes would need escaped. Not all languages will support the same escape method, so I recommend it being set by an optional parameter.
Example of what will fail:
Name
-----------
Fry's
INSERT INTO [Name] Values ( 'Fry's' )
Corrected for quotes (such as needing to escape a single quote with an additional single quote)
INSERT INTO [Name] Values ('Fry''s')
Example code:
$ConvertToSqlParams = @{TableName = ‘dbo.Names’
Path = ‘C:\temp\data.xlsx’
ConvertEmptyStringsToNull = $true
UseMSSQLSyntax = $true
SingleQuoteStyle = "''"
}
$SQLInsert = ConvertFrom-ExcelToSqlInsert @ConvertToSqlParams
Move code into the function Get-PropertyNames and remove the rest.
Now it only replaces $Columns with $ImportColumns after a couple of checks. So the heavy work like arranging and getting the right values is done in the original way of Import-Excel.
