Author Posts

November 19, 2016 at 8:04 pm

I have a script that converts a CSV file to an ASC formatted file with the data in specific column positions.

I am trying to take a Sales value in my array that has way too many trailing decimal places and round it to 2 decimal places and then I need it padded as well. The Sales value in the array is formatted like "354.9999999999" or "21.98" or "967.000001" etc. It's not consistent. There are no commas in it.

Below is the existing old code where it's currently just padding the Sales field. I need it first round the sales number and then pad to 2 decimal places.

    $array = Import-csv -path $FileName.FullName  -delimiter ',' -Header RecordType,Date,Time,Code,Sales,Customers,Items
    $arr = $array | foreach-object {$_.RecordType + ' ' + $_.Date + ' ' + $_.Time + '    ' + $_.Code.padleft(5,' ') + '   ' +$_.Sales.padleft(10,' ') + $_.Customers.padleft(5,' ') + $_.Items.padleft(5,' ') }

November 19, 2016 at 11:16 pm

I assume you're looking at something like this:

$values = "354.9999999999","21.98","967.000001"
foreach ($value in $values) {

If that is the case, there is an easier way than looping using calculated expressions:

$array = Import-csv -path $FileName.FullName  -delimiter ',' -Header RecordType,Date,Time,Code,Sales,Customers,Items
$array | Select RecordType,

November 20, 2016 at 12:35 am

Hello to both, neither of these solutions helped/worked. Let me clarify a bit more of what I am trying to do.

I have a CSV file that is being converted to ASC format by way of this powershell script. The CSV file contains the following columns and formats.

RecordType = the letter "D"
Time = HHMM in military time format no colon
Code = 5 digits of text
Sales = number with decimals sometimes, sometimes not. Examples of values include (4.0099999999999998, 31.02, 85, 3.5800000000000001)
Customers = number with no decimals
Items = number with no decimals

Sample of the CSV:

Record Type,Date,Time,Code,Sales,Customers,Items

The Sales column is what is causing me so much trouble. I need the trailing decimals to not be there, it needs to be only 2 decimal places.

The output once in ASC format needs to be in a very specific column layout, hence the padding I have in my script).

A sample ASC output format should look like this:

D 20151107 0630    FDSVC        10.76    1    4
D 20151107 0945    FDSVC         5.46    3    4

I have never coded in Powershell before, I am simply trying to update the existing script that I have to try and get the Sales number to round AND padleft.

Your help is very much appreciated!

November 20, 2016 at 2:50 am

First: You should have more faith in us! 😉 😀
Second: You really should start to learn Powershell. I am sure it will pay off for you in the future. 😉

$Array = Import-Csv -Path -place your path here-\sample.csv -Delimiter ','
$Array | 
Foreach-Object -Process { "{0,1} {1,7} {2,4} {3,8} {4,12:n2} {5,4} {6,4}" -f $_.'Record Type',$_.Date,$_.Time,$_.Code,[Math]::Round($_.Sales,2),$_.Customers,$_.Items} | 
Out-File -FilePath -place your path here-\asc.asc -Encoding utf8 -Append -Force

November 21, 2016 at 12:56 pm

Much easier to help with examples and expected output. 🙂

November 21, 2016 at 2:21 pm

Thanks for the help, unfortunately it's still not working. Here is what it is outputting....

D 20151106 0215    MERCH   4.0099999999999998  
D 20151106 0515    FDSVC   1.1899999999999999  
D 20151106 0530    FDSVC   1.5900000000000001  
D 20151106 0600    FDSVC   0.98999999999999999 
D 20151106 0615    FDSVC        21.18    8   13
D 20151106 0615    MERCH        22.02    5    6

Also, if it helps you out, I've pasted the entire larger script code below. There is a lot of other stuff it does so I initially didn't paste it all as to not overload you with info.

if (Get-Variable RootDir -Scope Global -ErrorAction SilentlyContinue)
  $FromMainEpos = 'Y'
  write-host 'Reading Config file'
  $MyRunPath=Split-Path $script:MyInvocation.MyCommand.Path
  $scriptPath = "$MyRunPath\ReadConfig.ps1"
  Invoke-Expression "$scriptPath"
  $sftplocation     = "$RootDir\Scripts\temp\tempholding"  
  $EPOSdir 		    = "$RootDir\Data"
  $logdir           = "$RootDir\Scripts\logs"
  $logfile          = "$logdir\log-overall.txt"

# General Variables
    $date 			= get-date
    $scriptname 	= $MyInvocation.MyCommand.Name
    $servername		= get-content env:computername

    ## Splunk Logging variables
    # $MyAccount - take from teh Configuration file	
    # Set the datestamp just before each log message
    $MyModule='Auto Down Load'
	$masterlog = "$RootDir\logs\eposlog.txt"

# return a result of 0 to indicate all extra process was successful 
# return a result of 1 to force the whole process to exit
$result = 0

$SourcePath = "$RootDir\Scripts\temp\tempholding"
$TargetPath = "$RootDir\Scripts\Temp\PointofSales" 

$MyDateTimeStamp = Get-Date -f "yyyy-MM-dd HH:mm:ss,fff"
$MyMessage='SFTP CSV to ASC conversion Starting'
$LogMessage = $MyDateTimeStamp + ' | ' + $MyPriority.padright(6).substring(0,6) + ' | ' + $MyAccount.padright(10).substring(0,10) + ' | ' + $MySystem.padright(3).substring(0,3) + ' | ' + $MyModule.padright(20).substring(0,20) + ' | ' + $MyMessage
Add-Content $logfile  $LogMessage

$Prefix = ""
$FilterVal   = $Prefix + '*.csv'
$CSVFiles = Get-ChildItem -Path $SourcePath -Filter $FilterVal | sort -unique
ForEach ($FileName in $CSVFiles)
  if ($FileName.FullName  -ne $null)
    $array = Import-csv -path $FileName.FullName  -delimiter ',' -Header RecordType,Date,Time,Code,Sales,Customers,Items
	$arr = $array | Foreach-Object -Process { "{0,1} {1,7} {2,4} {3,8} {4,12:n2} {5,4} {6,4}" -f $_.'Record Type',$_.Date,$_.Time,$_.Code,[Math]::Round($_.Sales,2),$_.Customers,$_.Items}
    # $arr = $array | foreach-object {$_.RecordType + ' ' + $_.Date + ' ' + $_.Time + '    ' + $_.Code.padleft(5,' ') + '   ' + [math]::Round($_.Sales,2).PadLeft(10,' ') + $_.Customers.padleft(5,' ') + $_.Items.padleft(5,' ') }
	$BaseName = $FileName -replace ".csv", ""
    $OutputFile = $TargetPath + '\' + $BaseName + '.asc'
    # $MyDateTimeStamp = Get-Date -f "yyyy-MM-dd HH:mm:ss,fff"
    # $MyMessage='Input File = ' $FileName.FullName    'Output File = '  $OutputFile
    # $MyPriority='INFO'
    # $LogMessage = $MyDateTimeStamp + ' | ' + $MyPriority.padright(6).substring(0,6) + ' | ' + $MyAccount.padright(10).substring(0,10) + ' | ' + $MySystem.padright(3).substring(0,3) + ' | ' + $MyModule.padright(20).substring(0,20) + ' | ' + $MyMessage
    # Add-Content $logfile  $LogMessage	
    foreach ($i in $arr)
      $Record = $i.substring(0,47)
	  If ($SkipLine -eq 'N')
        Add-Content $OutputFile $Record

$MyDateTimeStamp = Get-Date -f "yyyy-MM-dd HH:mm:ss,fff"
$MyMessage='SFTP CSV to ASC conversion complete'
$LogMessage = $MyDateTimeStamp + ' | ' + $MyPriority.padright(6).substring(0,6) + ' | ' + $MyAccount.padright(10).substring(0,10) + ' | ' + $MySystem.padright(3).substring(0,3) + ' | ' + $MyModule.padright(20).substring(0,20) + ' | ' + $MyMessage
Add-Content $logfile  $LogMessage


return $result

November 21, 2016 at 3:02 pm

In my experience it helps usually debugging a large script to split it in smaler but still functional chunks. Try to run only the part you asked here to put in a seperate script and see if it works. If not you might have some inconsistencies in your csv file.

I've seen you import your csv and put some extra headers. Doesn't your csv already have some? Also check if the header names are accordingly named ('RecordType' vs 'Record Type').

November 21, 2016 at 3:12 pm

Getting closer!

It converted properly once I fixed the "Record Type" vs "RecordType" header issue.

But the first row of data in the file is getting dropped. Powershell is throwing this error...

Cannot convert argument "0", with value: "Sales", for "Round" to type "System.Double": "Cannot convert value "Sales" to
type "System.Double". Error: "Input string was not in a correct format.""
At .ps1:63 char:148
+ $arr = $array | Foreach-Object -Process { "{0,1} {1,7} {2,4} {3,8} {4,12:n2} {5,4} {6,4}" -f $_.'RecordType',$_.D
ate,$_.Time,$_.Code,[Math]::Round <<<< ($_.Sales,2),$_.Customers,$_.Items} + CategoryInfo : NotSpecified: (:) [], MethodException + FullyQualifiedErrorId : MethodArgumentConversionInvalidCastArgument

November 21, 2016 at 3:16 pm

I didn't dig too deep into your script, but I don't see where you clear your output file if it already exists. If you run this multiple times, it will keep accumulating records, which seems to be the case, since some records are in the old format.

if (test-path $OutputFile) {clear-content $OutputFile}

November 21, 2016 at 3:19 pm

Your input file already has headers and you are creating new ones.

    Specifies an alternate column header row for the imported file. The column header determines the names of the properties of the object that Import-CSV crea

    Enter a comma-separated list of the column headers. Enclose each item in quotation marks (single or double). Do not enclose the header string in quotation
    marks. If you enter fewer column headers than there are columns, the remaining columns will have no header. If you enter more headers than there are column
    s, the extra headers are ignored.

    When using the Header parameter, delete the original header row from the CSV file. Otherwise, Import-CSV creates an extra object from the items in the head
    er row.

    Required?                    false
    Position?                    named
    Default value
    Accept pipeline input?       false
    Accept wildcard characters?  false

November 21, 2016 at 3:20 pm

The input file name is always a unique file name so I don't need to worry about clearing the output file.

November 21, 2016 at 3:23 pm

If your csv file originally has headers you don't have to add them while you import it. Otherwise it is necessary. You just posted in your csv sample some lines and rows including headers and in your script you added headers 'again' ... that's what I meant.

November 21, 2016 at 3:41 pm

The CSV being imported will always have these headers in the input file. How do I exclude them then in my code?

November 21, 2016 at 3:45 pm

There's no reason to rename them unless you suspect they might change (but not the data layout). Just remove the -header parameter and go back to the original name for 'Record type'. If, for some reason you felt you had to rename them, then just skip the first record.

$array[1..($array.count – 1)] | ...

November 21, 2016 at 3:53 pm

Why don't you play a little bit around in the console until you understand exactly what's going on? In the console you have immidiate response and you can see how it works.

Get-Help Import-CSV -Full

Try to compare this:

Import-CSV -Path 'Your file path here' 

with this:

Import-CSV -Path 'Your file path here' -Header 'Your headers here'

... you know what I mean?

BTW: You can debug your script line by line in the ISE.

November 21, 2016 at 3:53 pm

You can also get rid of the skip line logic in your output. They headers won't be in your output array.

November 21, 2016 at 3:54 pm

I am not sure what you mean by this.... $array[1..($array.count – 1)] | ...

I'm a total newbie here. 🙂

Below is the code as it stands based on your recommendation to remove the -header info. The scripts runs without error now, however the first row of actual data from the CSV input file is missing from the output ASC file.

**********EDIT: I changed the skip line parameter to = 'N' and all is working now!!!*****************

$array = Import-csv -path $FileName.FullName  -delimiter ',' 
	$arr = $array | Foreach-Object -Process { "{0,1} {1,7} {2,4} {3,8} {4,12:n2} {5,4} {6,4}" -f $_.'Record Type',$_.Date,$_.Time,$_.Code,[Math]::Round($_.Sales,2),$_.Customers,$_.Items}

Input data:

Record Type,Date,Time,Code,Sales,Customers,Items

Output ASC file:

D 20151106 0515    FDSVC         1.19    1    1
D 20151106 0530    FDSVC         1.59    1    1

November 21, 2016 at 4:00 pm

I'm a total newbie here. 🙂

Maybe you should take a little time and start to learn Powershell 'from scratch'. So you understand better how to debug and how to get help and information you need. Here you can find some great recourses to start from: Beginner Sites and Tutorials

November 21, 2016 at 4:00 pm

Thank you ALL so much for your help on this! You have no idea how much I appreciate this!

November 21, 2016 at 4:00 pm

Sometimes there's no getting around reading the help files. 🙂

See: help about_array

What I did was index the array from the 2nd line (base 0 array) to the last line.

But, all of that is unnecessary unless you need to mess with the headers, which seems unlikely. When you use the existing headers, they won't be in the input array or the output array.