Developer42

2015-12-03

SQL Server :: Getting Column Length of Various Columns

Filed under: Microsoft, SQL Server, Technology — Tags: , , , , — Developer42 @ 19:17

Sometimes we need to lookup the maximum length allowed for various columns in a number of tables.
This script allows you to specify a list of table/column pairs into the table variable @interesting, then pulls back the max_length information for these, taking into account the data type (i.e. so data types storing 2 byte character sets can cope predictably).

It works by creating a temp table with a column holding a single character of each (string) data type, then reading the length of that field to get the length of a single character of that data type, and dividing the max length of the fields we’re after by this multiplier to get the length in characters rather than bytes.


declare @interesting table(col sysname, tab sysname)
insert @interesting
values ('someColumn','someTable')
,('secondColumn','someTable')
,('column3','anotherTable')

declare @JBDataTypes table(system_type_id int, user_type_id int, max_length int)
declare @sql nvarchar(max) 
select @sql = coalesce(@sql + ',','create table #dataTypesJB (') + quotename(name) + ' ' + name  
from sys.types 
where collation_name is not null --string fields only
set @sql = @sql + '); 
select system_type_id
, user_type_id 
, max_length 
from tempdb.sys.columns 
where object_id = object_id(''tempdb..#dataTypesJB''); 
drop table #dataTypesJB'
insert into @JBDataTypes exec (@sql)

select c.max_length / coalesce(jb.max_length,1) max_length 
, i.col
, i.tab
from @interesting i
inner join sys.columns c 
on c.object_id = object_id(i.tab)
and c.name = i.col
left outer join @JBDataTypes jb
on jb.system_type_id = c.system_type_id
and jb.user_type_id = c.user_type_id 
order by i.tab,i.col

2015-05-18

Microsoft Dynamics AX 2012 Chart of Accounts Advanced Validation Rules SQL

A follow up to an earlier article: Microsoft Dynamics AX 2012 Chart of Accounts Validation Rules SQL.

The Advanced Rules can be found in AX under: General Ledger > Setup > Chart of accounts > Configure account structures > select account structure > Advanced rule.
Here’s a small sample of our advanced rules from AX:

Advanced Rules in AX

Advanced Rules in AX

Here’s the same rule (results row 9), as produced by the below SQL:

AX Advanced Rules in SQL

AX Advanced Rules in SQL

Full SQL:

select dr.name AdvancedRule
, dr.description Name
, a.Name AccountStructure
, a.Description ASDescription
, case when dr.IsDraft =1 then 'Draft' else 'Active' end [Status] --seems to just duplicate the IsDraft info
, dr.IsDraft Draft
, drc.Filter
, dh.Name ARSAdvancedRuleStructure
, dh.DESCRIPTION ARSName
from DimensionRule dr
inner join DimensionHierarchy a on a.recid = dr.AccountStructure and a.Partition = dr.Partition
inner join (
       select drcO.DimensionRule
       , drcO.Partition
       , stuff
       (
              (
                     select ' And ' + da.Name + ' '
                     + case
                           when coalesce(drcI.WildCardString,'') > ''
                                  then case
                                         when drcI.WildCardString like '\%%\%' escape '\' then 'contains '
                                         when drcI.WildCardString like '%\%' escape '\' then 'begins with '
                                         when drcI.WildCardString like '\%%' escape '\' then 'ends with '
                                         else 'equals '
                                  end + quotename(replace(drcI.WildCardString,'%',''),'''')
                           when drcI.IsFromOpen = drcI.IsToOpen then --if both are 0 or both are 1 (second scenario seems weird, but that's how the app behaves)
                                  case
                                         when drcI.RangeFrom = drcI.RangeTo then 'Is ' + quotename(drcI.RangeFrom,'''')
                                         else 'Between ' + quotename(drcI.RangeFrom,'''') + ' through ' + quotename(drcI.RangeTo,'''')
                                  end
                                            --in the below statements IsFromOpen and IsToOpen seem to behave backwards to what you'd expect; but again that's what the app does
                           when drcI.IsFromOpen=1 and drcI.IsToOpen=0 then 'Greater than or equal to ' + quotename(drcI.RangeFrom,'''')
                           when drcI.IsFromOpen=0 and drcI.IsToOpen=1 then 'Less than or equal to ' + quotename(drcI.RangeTo,'''')
                           else '-UNEXPECTED SCENARIO; SPEAK TO JB-' --this should nevere happen
                     end 
                     FROM DimensionRuleCriteria drcI
                     inner join DimensionAttribute da on da.RecId = drcI.DimensionAttribute and da.Partition = drcI.Partition
                     where drcI.DimensionRule = drcO.DimensionRule and drcI.Partition = drcO.Partition
                     order by da.Name --drcI.RecId
                     for xml path(''), type
              ).value('.','nvarchar(max)')
       ,1,4,'Where') Filter
       FROM DimensionRuleCriteria drcO
       group by drcO.Partition
       , drcO.DimensionRule
) drc on drc.DimensionRule = dr.RecId and drc.Partition = dr.Partition
inner join DimensionRuleAppliedHierarchy drah on drah.DimensionRule = dr.RecId and drah.Partition = dr.Partition
inner join DimensionHierarchy dh on dh.recid = drah.DimensionHierarchy and dh.Partition = drah.Partition
order by a.Name, dr.name, dh.Name, dr.IsDraft

NB: One anomaly you may notice is some rules appear twice; once as Draft and once as Active. This is where an existing rule is being edited; the existing rule remains active whilst its replacement is created in draft status. When you browse the rules you’ll see the draft rule; however the active rule is the one being applied in any validation.

2015-03-09

Microsoft Dynamics AX 2012 Chart of Accounts Validation Rules SQL

In doing some analysis work on performance, a consultant recently asked for an extract of all of our AX2012’s COA validation rules.
The below SQL is an attempt at regenerating the rules based off the data in the various related tables.
NB: I’m not 100% certain that this gives all rules as they’d appear in the application, but a few initial tests prove promising; so hopefully this SQL can be reused by others with similar requirements.

Here’s a small sample of our validation rules from AX:

AX 2012 Chart Of Accounts Validation Rules Screenshot

AX 2012 Chart Of Accounts Validation Rules Screenshot

Here’s the same rules, as produced by the below SQL:

AX 2012 Chart Of Accounts Validation Rules SQL Results Screenshot

AX 2012 Chart Of Accounts Validation Rules SQL Results Screenshot

Full SQL:

set transaction isolation level read uncommitted
go

;with c1 as
(
       select DimensionConstraintNode
       , STUFF
       (
              (
                     SELECT ';' + case 
                           when WILDCARDSTRING > '' then WILDCARDSTRING 
                           when RangeFrom=RangeTo then RangeFrom 
                           else concat(RangeFrom,'..',RangeTo) 
                     end 
                     FROM DimensionConstraintNodeCriteria inside
                     where inside.DimensionConstraintNode = outside.DimensionConstraintNode
                     ORDER BY ORDINAL, RECID
                     FOR XML PATH (''),TYPE
              ).value('.','nvarchar(max)')
              ,1,1,''
       ) AS Rules
       from DimensionConstraintNodeCriteria outside
       group by DimensionConstraintNode
)
, ActiveDCN as
(
       select dct.dimensionhierarchy 
       , dcn.DIMENSIONCONSTRAINTTREE
       , dhl.level_
       , dhl.dimensionattribute
       , da.name AttributeName
       , dcn.dimensionhierarchylevel
       , dcn.PARENTCONSTRAINTNODE
       , dcn.RecId
       , c1.Rules
       from DimensionConstraintTree dct 
       inner join DIMENSIONHIERARCHYLEVEL dhl on dhl.dimensionhierarchy = dct.dimensionhierarchy
       inner join DIMENSIONATTRIBUTE da on da.recid = dhl.dimensionattribute 
       inner join DimensionConstraintNode dcn on dcn.DIMENSIONCONSTRAINTTREE = dct.recid and dcn.dimensionhierarchylevel = dhl.RECID
       left outer join c1
       on c1.DimensionConstraintNode = dcn.recid
    where (dcn.activeto = '1900-01-01 00:00' or dcn.activeto >= getutcdate())
    and (dcn.activefrom = '1900-01-01 00:00' or dcn.activefrom <= getutcdate())
)
select dh.name HierarchyName
, concat(dcn1.AttributeName,': ',coalesce(dcn1.Rules,'*')) R1
, concat(dcn2.AttributeName,': ',coalesce(dcn2.Rules,'*')) R2
, concat(dcn3.AttributeName,': ',coalesce(dcn3.Rules,'*')) R3
, concat(dcn4.AttributeName,': ',coalesce(dcn4.Rules,'*')) R4
from DIMENSIONHIERARCHY dh
inner join ActiveDCN dcn1 on dcn1.dimensionhierarchy = dh.recid and dcn1.level_ = 1 --not sure if level’s required; it does affect results; keeping in for now as improves performance (though also reduces result set by 75%)
left outer join ActiveDCN dcn2 on dcn2.ParentConstraintNode = dcn1.Recid --and dcn2.level_ = 2
left outer join ActiveDCN dcn3 on dcn3.ParentConstraintNode = dcn2.Recid --and dcn2.level_ = 3
left outer join ActiveDCN dcn4 on dcn4.ParentConstraintNode = dcn3.Recid --and dcn2.level_ = 4
--where dcn1.Rules like '%640103%' --to produce screenshot
order by dh.name, dcn1.DIMENSIONCONSTRAINTTREE

2014-09-26

PowerShell Script :: Get AD Users by Email (Advanced)

Filed under: Microsoft, Technology — Tags: , , , , , , , — Developer42 @ 22:19

Here’s a script I knocked up today.

Auto.ps1 allows me to host the script on a server (or wherever), whilst others can use it by dropping input files into a queue folder, without needing to touch powershell (which may be scary to non-programmers, or may require additional setup or permissions).
ADGetUsersByEmailAdvanced.ps1 gets AD info based on email addresses; without requiring exchange modules, and includes workarounds to cope with missing information.

Auto.ps1

Monitors a folder for new text files. Once found, passes that file to a script to be processed. On completion moves the source file to the same directory & renames to begin with the same timestamp as the generated output file.

$infile = "\\myServer\myShare\Scripts\powershell\ADGetUsersByEmail\in\*.txt"

while (1 -eq 1) {
	#wait for a new file
	while(!(Test-Path $infile)) {Start-Sleep -s 30;}
	write-host "new input file found"
	Get-ChildItem $infile | %{
		$fileTimestamp = "{0:yyyy-MM-dd_HHmmss}" -f (get-date).ToUniversalTime()
		$inputFile = $_.fullname
		$exportFile = "{0}\out\{1}_output.csv" -f $PSScriptRoot,$fileTimestamp
		$inputFileMoved = "{0}\out\{1}_{2}" -f $PSScriptRoot,$fileTimestamp,$_.name
		write-host ("source: {0}" -f $inputFile)
		write-host ("output: {0}" -f $exportFile) 
		write-host ("archive: {0}" -f $inputFileMoved) 
		.\ADGetUsersByEmailAdvanced.ps1 -sourceFile $_.fullname -exportFile $exportFile 
		write-host "processed"
		Move-Item $inputFile $inputFileMoved
		write-host "archived"
	}
}

ADGetUsersByEmailAdvanced.ps1

Given a text file containing a list of email addresses, attempts to resolve those to corresponding AD users, taking advantage of email information in AD where available, then gracefully degrading to more hacky methods. Works its way through a list of domains in case the users are in the same company but on a different domain.

Param(
  [string]$sourceFile
  ,[string]$exportFile 
)
#$sourceFile = '.\sourceEmails.txt'
#$exportFile = ".\output_{0:yyyy-MM-dd_HHmmss}.csv" -f (get-date).ToUniversalTime()
$domains = 'eu','usa','myDomain','anotherDomain' # domain points to the GC; could equally list GC server names here, though this version's more user friendly

#create dummy; allows us to put in values for any unfound items (currently just using null, but we can easily amend if desired)
$adDummy = New-Object –TypeName PSObject –Prop @{
	emailSearched	= $null;
	notFound		= $true;
	sAmAccountName 	= $null;
	fullname		= $null;
	firstname		= $null;
	lastname		= $null;
	cn				= $null;
	countryCode		= $null;
	country			= $null;
	#title			= $null;
	title			= $null;
	department		= $null;
	company			= $null;
	email			= $null;	
	adEmail			= $null;
	proxyEmail		= $null;
}

function RemoveEmailDomain($email) {
  return $email -replace "(\S*)@\S*", '$1'
}
function IsFirstDotLast($name) {
	return $name -like '*.*'
}
function GetFirstName($name) {
	return $name -replace "(\S*?)\.\S*", '$1'
}
function GetLastName($name) {
	return $name -replace "\S*?\.(\S*)", '$1'
}
function GetFirstNamePartial($name) {
	return $name.substring(0,[system.math]::min(3,$name.length))
}
function GetLastNamePartial($name) {
	return $name.substring([system.math]::max($name.length-3,0),[system.math]::min(3,$name.length))
}
function GetADUserByIdentity($id, $domain) {
	#trycatch since erroraction not recognised on this type of command & I don't want error messages polluting my output
	write-host "searching for id '$id' on domain '$domain'"
	try { 
		Get-ADUser -Identity $id -Server $domain -Properties sAmAccountName, displayName, givenName, surname, distinguishedName, countryCode, c, title, department, company, emailAddress, proxyAddresses
	} catch {}
}
function GetADUserFiltered($filter, $domain) {
	Get-ADUser -Filter $filter -Server $domain -Properties sAmAccountName, displayName, givenName, surname, distinguishedName, countryCode, c, title, department, company, emailAddress, proxyAddresses
}
function GetADUserByEmailAddress($email, $domain) {
	write-host "searching for email '$email' on domain '$domain'"
	$filter = {emailAddress -eq $email} 
	GetADUserFiltered $filter $domain
}
function GetADUserByProxyAddress($email, $domain) {
	write-host "searching for proxy '$email' on domain '$domain'"
	$psBugFixSearchUser = "*:$_*"
	$filter = {proxyAddresses -like $psBugFixSearchUser}
	GetADUserFiltered $filter $domain
}
function GetADUserByFullName($name, $domain) {
	$fn = GetFirstName $name
	$ln = GetLastName $name
	write-host "searching for name '$fn', '$ln' on domain '$domain'"
	$filter = {(givenname -eq $fn) -and (surname -eq $ln)}
	GetADUserFiltered $filter $domain
}
function GetADUserByPartialName($name, $domain) {
	$fn = "{0}*" -f (GetFirstNamePartial $name)
	$ln = "*{0}" -f (GetLastNamePartial $name)
	write-host "searching for partial name '$fn', '$ln' on domain '$domain'"
	$filter = {(givenname -like $fn) -and (surname -like $ln)}
	GetADUserFiltered $filter $domain | where { ($_.givenname + $_.surname) -eq $name }
}
function FindBestMatch($email) {
	$result = $null
	#$domains | %{ $result=GetADUserByEmailAddress $email $_; if($result) {return $result;} } #doesn't play as expected
	foreach($domain in $domains) { $result=GetADUserByEmailAddress $email $domain; if($result) {return $result;} }
	foreach($domain in $domains) { $result=GetADUserByProxyAddress $email $domain; if($result) {return $result;} }
	$name = RemoveEmailDomain $email
	foreach($domain in $domains) { $result=GetADUserByIdentity $name $domain; if($result) {return $result;} }
	if(IsFirstDotLast($name)) {
		foreach($domain in $domains) { $result=GetADUserByFullName $name $domain; if($result) {return $result;} }
	} else {
		foreach($domain in $domains) { $result=GetADUserByPartialName $name $domain; if($result) {return $result;} }
	}	
	return $adDummy;
}

#define a function for later use
#get list of emails (ignore blanks)
$emails = (Get-Content $sourceFile) | where{ $_ -gt '' }

#get data from ad and stick it in an csv (or error to console if not found)
$emails | %{ 
	#get ad user by email address
	$adUser = FindBestMatch($_);
	if($adUser.notFound) {
		write-host ":(" -ForegroundColor Red
	} else {
		write-host ":)" -ForegroundColor Green
	}
	#return object replresenting results.
	New-Object –TypeName PSObject –Prop @{
		emailSearched	= $_;
		found			= if($adUser.notFound){$false} else {$true};
		sAmAccountName 	= $adUser.sAmAccountName;
		fullname		= $adUser.displayName;
		firstname		= $adUser.givenname;
		lastname		= $adUser.surname;
		cn				= $adUser.distinguishedName;
		countryCode		= $adUser.countryCode;
		country			= $adUser.c;
		#title			= $adUser.personalTitle;
		title			= $adUser.title;
		department		= $adUser.department;
		company			= $adUser.company;
		adEmail			= $adUser.emailAddress;
		proxyEmail		= [string]$adUser.proxyAddresses; #string joins the array down to a single string value
	}
} | export-csv $exportFile -notype #stick output to file

Script could be improved by allowing auto to kick off jobs so multiple instances of the worker script can be run simultaneously. Also changing the main script to make use of workflows and take advantage of the parallel foreach method should significantly improve it’s performance. However I’m still pretty new to PowerShell, so those steps will have to come later.

Create a free website or blog at WordPress.com.

%d bloggers like this: