Transpose rows to columns in PowerShell - powershell

Question

I have a source file with the below contents:
0
ABC
1
181.12
2
05/07/16
4
Im4thData
5
hello
-1
0
XYZ
1
1333.21
2
02/02/16
3
Im3rdData
5
world
-1
...
The '-1' in above lists is the record separator which indicates the start of the next record. 0,1,2,3,4,5 etc are like column identifiers (or column names).
This is my code below.
$txt = Get-Content 'C:myfile.txt' | Out-String
$txt -split '(?m)^-1\r?\n' | ForEach-Object {
$arr = $_ -split '\r?\n'
$indexes = 1..$($arr.Count - 1) | Where-Object { ($_ % 2) -ne 0 }
$arr[$indexes] -join '|'
}
The above code creates output like below:
ABC|181.12|05/07/16|Im4thData|hello
XYZ|1333.21|02/02/16|Im3rdData|World
...
But I need output like below. When there are no columns in the source file, then their row data should have blank pipe line (||) like below in the output file. Please advise the change needed in the code.
ABC|181.12|05/07/16||Im4thData|hello ← There is no 3rd column in the source file. so blank pipe line (||).
XYZ|1333.21|02/02/16|Im3rdData||World ← There is no 4th column column in the source file. so blank pipe line (||).
...

Answer 1

If you know the maximum number of columns beforehand you could do something like this:
$cols = 6
$txt = Get-Content 'C:myfile.txt' | Out-String
$txt -split '(?m)^-1\r?\n' | ForEach-Object {
# initialize array of required size
$row = ,$null * $cols
$arr = $_ -split '\r?\n'
for ($n = 0; $n -lt $arr.Count; $n += 2) {
$i = [int]$arr[$n]
$row[$i] = $arr[$n+1]
}
$row -join '|'
}
Otherwise you could do something like this:
$txt = Get-Content 'C:myfile.txt' | Out-String
$txt -split '(?m)^-1\r?\n' | ForEach-Object {
# create empty array
$row = #()
$arr = $_ -split '\r?\n'
$k = 0
for ($n = 0; $n -lt $arr.Count; $n += 2) {
$i = [int]$arr[$n]
# if index from record ($i) is greater than current index ($k) append
# required number of empty fields
for ($j = $k; $j -lt $i-1; $j++) { $row += $null }
$row += $arr[$n+1]
$k = $i
}
$row -join '|'
}

Answer 2

Needs quite a bit of processing. There might be a more efficient way to do this, but the below does work.
$c = Get-Content ".\file.txt"
$rdata = #{}
$data = #()
$i = 0
# Parse the file into an array of key-value pairs
while ($i -lt $c.count) {
if($c[$i].trim() -eq '-1') {
$data += ,$rdata
$rdata = #{}
$i++
continue
}
$field = $c[$i].trim()
$value = $c[++$i].trim()
$rdata[$field] = $value
$i++
}
# Check if there are any missing values between 0 and the highest value and set to empty string if so
foreach ($row in $data) {
$top = [int]$($row.GetEnumerator() | Sort-Object Name -descending | select -First 1 -ExpandProperty Name)
for($i = 0; $i -lt $top; $i++) {
if ($row["$i"] -eq $null) {
$row["$i"] = ""
}
}
}
# Sort each hash by field order and join with pipe
$data | ForEach-Object { ($_.GetEnumerator() | Sort-Object -property Name | Select-Object -ExpandProperty Value) -join '|' }
In the while loop, we are just iterating over each line of the file. The field number an value are separated by a value of one, so each iteration we take both values and add them to the hash.
If we encounter -1 then we know we have a record separator, so add the hash to an array, reset it, bump the counter to the next record and continue to the next iteration.
Once we've collected everything we need to check if there are any missing field values, so we grab the highest number from each hash, loop over it from 0 and fill any missing values with an empty string.
Once that is done you can then iterate the array, sort each hash by field number and join the values.