parsing whitespace-separated fields - powershell

Question

Im trying to write a script which works as follows:
My input is a text file with 8 rows and 8 columns, filled with values 0 or 1, with a single space character each separating the columns.
I need to check the 4th number in each row, and output false, if it is 0, and true, if it is 1.
My code at the moment looks like this:
param($fname)
$rows = (Get-Content $fname)
for ($i=0;$i -lt $rows.Length;$i++)
{
if ($rows[$i][6] -eq 1)
{
Write-Host "true"
}
if ($rows[$i][6] -ne 1)
{
Write-Host "false"
}
}
So I use [$i][6], because I get that that's the 4th number, accounting for the number of spaces acting as separators.
I checked and thought it was perfect, but somehow it says false for every line, but when I Write-Host $rows[0][6] it is 1.

Answer 1

tl;dr
# Create sample input file, with values of interest in 4th field
# (0-based field index 3).
#'
0 0 0 0 0 0 0 0
0 0 0 1 0 0 0 0
'# > file
foreach ($line in (Get-Content file)) {
$fields = -split $line
if ($fields[3] -eq '1') {
"true"
} else {
"false"
}
}
yields:
false
true
There are many subtleties to consider in your original code, but the above code:
offers a more awk-like approach by splitting each input line into whitespace-separated fields, whatever the length of the fields, courtesy of the unary -split operator.
subscripts (indices) can then be based on field indices rather than character positions.
All fields returned by -split ... are strings, hence the comparison with string literal '1', but, generally, PowerShell performs a lot of behind-the-scenes conversion magic for you: with the code above - unlike with your own code - using 1 would have worked too.
As for why your approach failed:
Indexing into (using a subscript with) a string value in PowerShell is a special case: it implicitly treats the string as a character array, and, with a single index such as 6, returns a [char] instance.
It is the LHS (left-hand side) of an expression involving a binary operator such as -eq that determines what type the RHS (right-hand side) will be coerced to, if necessary, before applying the operator:
([char] '1') -eq 1 # !! $false
Coercing the (implied) [int] type of RHS 1 to the LHS type [char] yields Unicode codepoint U+0001, i.e., a control character rather than the "ASCII" digit '1', which is why the comparison fails.
#PetSerAl's helpful, but cryptic suggestion (in a comment on the question) to use '1'[0] rather than 1 as the RHS solves the problem in this particular case, because '1'[0] returns 1 as a [char] instance, but the solution doesn't generalize to multi-character field values.
'1' -eq 1 # $true; same as: ([string] 1) -eq 1 or ([string] 1) -eq '1'
Converting integer 1 to a string indeed is the same as '1'.

Answer 2

This script fills a true 2d array with the proper values from the matrix file, but the output doesn't fit.
$Array = New-Object "bool[,]" 8,8
[int]$i=0 ; [int]$j=0
get-content $fname|foreach{ foreach ($b in ($_ -split(' '))) {
"{0},{1}={2}" -f $i,$j,($b -eq 0)
$Array[$i,$j] = ($b -eq 0)
$j++}
$j=0; $i++}