Reporting Services dynamic multi-column parameter box

The Reporting Services parameter section at the top of a report is generated as a HTML table with two fixed columns. To display SSRS parameters in three columns in the parameter field is not possible using a configuration option or even a CSS style sheet, so element conversion is needed.

The instructions below fixes this for the /reports and /reportserver web entry points and are suited for Reporting Services 2012, but might work for earlier versions as well.

The code:

• Adds the jquery library to every report
• Uses this library to:
• Create a DIV element for each set of TD parameter elements (=label and textbox/dropdown) and add it to the first TD element of the first TR row
• Add all sets of TD parameters to these DIVs
• Remove all other TR rows from the parameter TABLE

The result is that as many parameters as possible are fitted on one horizontal row:

The code has to be appended to the end of the ReportingServices.js file. This file is added automatically to every report generated by Reporting Services.

The file is most probably located in the public javascript folder: c:\Program Files\Microsoft SQL Server\MSRS11.MSSQLSERVER\Reporting Services\ReportManager\js

Perform the following steps to get a multiple columns in the parameter box:

1) Download and add the publicly available jquery-1.4.4.min.js to the public javascript folder.

2) Code to append:

// helper function to add scripts to head of html
function addHeadElement(headtype, file) {
var head = document.getElementsByTagName('head')[0];
  if (headtype == 'script')
  {
    var script = document.createElement('script');
    script.setAttribute('type', 'text/javascript');
    script.setAttribute('src', file);
    head.appendChild(script); 
  }
}
addHeadElement('script','/Reports/js/jquery-1.4.4.min.js');

// function that converts the parameter table to divs 
function param() {

  // create as many div->table->tr elements in the first column of the first row as there are parameters
  $(".ParamLabelCell").each( function (i, v) {
    var html = "<div style='float: left; padding-right: 10px;'> \
      <table> \
        <tr id='blackninja" + i + "'></tr> \
      </table> \
    </div>";
    $("td.InterParamPadding").filter(":first").append(html);
  });
  
  // move each label to its proper div->table->tr element
  $("td.ParamLabelCell").each( function (i, v) {
    $(this).appendTo("#blackninja" + i);
  });
  
  // move each prompt to its proper div->table->tr element
  $("td.ParamEntryCell").each( function (i, v) {
    $(this).appendTo("#blackninja" + i);
  });
  
  // remove all other parameter rows
  $("tr[isparameterrow='true']").not(':first').remove();

}
// add the convert function to the onload event twice
// this guarantees a cross-browser safe implementation
function pageLoad() {
  param();
}
window[ addEventListener ? 'addEventListener' : 'attachEvent' ]( addEventListener ? 'load' : 'onload', param );

3) Change style sheet

Add the following code to the styles\ReportingServices.css style sheet to make the dropdown boxes and textboxes smaller in width. More parameters will fit on a row:

.ParametersFrame.ParamsGrid.MenuBarBkGnd { padding: 0px; }
.DisabledTextBox { width: 125px; }
td.ParamEntryCell [id$="Value"] { width: 125px; }
td.ParamEntryCell { padding: 0px; }
td.SubmitButtonCell { padding-top: 2px; padding-bottom: 0px; }
td.ParamLabelCell { padding-right: 5px; }
td.InterParamPadding { padding-left: 0px; }
/* IE11 Edge/default profile fix */
tr.MenuBarBkGnd { height: 0px; } 
#ParametersRowctl31 { height: 0px; }

4) Enable in the reportserver

Add the following two elements to the head-element in \ReportServer\Pages\ReportViewer.aspx:

<meta http-equiv="X-UA-Compatible" content="IE=9">
<link href="/Reports/styles/ReportingServices.css" type="text/css" rel="stylesheet">
<script type="text/javascript" src="/Reports/js/ReportingServices.js"></script>
<script type="text/javascript" src="/Reports/js/jquery-1.4.4.min.js"></script>

Reinstall OEM-installed Windows 8.1

One of the first steps of the re-installation process of Windows 8.1 is that it asks for a Windows key. There is a "bug" that prevents a valid Windows key from working when the general Windows 8 installation disk is used on a laptop with a pre-installed OEM version.

Fortunately, if there are no installation disks supplied, then Windows 8.1 can still be easily re-installed on such a laptop, for example on another partition or disk:

1. Get a copy of the Windows 8.1 ISO, e.g. "Windows_8.1_EN-US_x64.ISO" and extract the contents to a USB stick

2. Bypass the Windows key screen by creating an ei.cfg text file in the sources folder on the USB stick. It should contain the following (replace "Core" with "Professional", if you have a valid key for that edition):

[EditionID]

Core

[Channel]

Retail

[VL]

0

3. Restart the laptop with the USB stick in it. The initial Windows key screen is skipped. During the latter part of the installation another Windows key screen pops up. Here the genuine valid key can be entered

Notes

• It has to be a USB stick, an external disk will not work: Win 8 can only be installed from an USB stick or a DVD
• Unfortunately, I cannot supply a link to the above mentioned ISO file
• Make sure the BIOS settings are such that the USB stick can be booted from and is the first in the sequence of bootable devices. Check how to enter the BIOS and set the bootable USB device on the site of the laptop manufacturer. Each manufacturer has a different method. For Lenovo IdeaPad: press and hold the small round button next to the power adapter entrance during a reboot

SQL Server contained database feature

SQL Server 2012 has a "contained database" feature.

In the normal case the logins of the (SQL Server) server are coupled to the users in each of its databases. The logins are stored in its MASTER database. If a database gets restored to another server, then the mapping is lost and you end up with "orphaned users". Those accounts can no longer access the database.

For a contained database on a SQL Server:

• The server logins are not used/required. Users can be created and used without a matching login on the server. These users are called "portable users"
• The database handles the autorization instead of the server. This is called "contained database authentication". Activate this delegation of functionality on the server:
  

  EXEC sp_configure 'contained database authentication',1
  GO
  RECONFIGURE

  In SSMS, a new option appears in the ->Security->Users->New User... screen: "SQL User with password"
• Works for SQL and Active Directory Accounts, e.g. statement for both without logins:
  

  CREATE USER [testdomain\testuser]
  CREATE USER [biuser]

Notes:

• Partial containment: in 2012 only the "partial containment" functionality works. In the future a "full containment" option if foreseen
• Connect from SSMS: you need to specify the database, otherwise the connection attempt fails, because the database is now doing the authentication
• Security: once a portable user gains access to a contained database through contained database authentication, that user also ends up gaining guest access to all other databases on the host system
• Conversion: conventional existing users can be converted into portable users through the use of a special stored procedure: sp_migrate_user_to_contained. It provides an argument that you can use to specify whether or not to disable the server-level login -- something you'll typically want to do as a best practice to avoid ugly login problems that can occur when duplicate logins and users overlap each other
• Downsides: no cross-database joins, linked server use, database mail, possible tempdb colation issues. Use "select * from  sys.dm_db_uncontained_entities" to check containment readiness

PowerPivot cannot load from procedure with info messages

Data can be loaded into PowerPivot from a SQL Server stored procedure, but the procedure cannot contain:

• PRINT statements, like print 'lets start'
• informational RAISERROR events, like raiserror ('lets start', 0, 1)

Both return messages to the client. Tools like Management Studio can handle these messages and display them in a separate window. PowerPivot handles them like the returned data and fails reporting an error. The error is not very helpful:

OLE DB or ODBC error: lets start; 01000.
An error occurred while processing table 'Query'.
The current operation was cancelled because another operation in the transaction failed.

The only solution is to remove these instructions from the procedure and any procedures it might call. There is no way to suppress this on the OLEDB connection during the call to the procedure.

It is also advisable to have the following two instructions at the start of the procedure:

• set fmtonly off
• set nocount on

LISTAGG / implode solution for SQL Server

Group multiple records into one and concatenate the values of the string/text values of a field:

ALTER view [dbo].[vcomments] as
with x as
(
select
cmt_domain,
cmt_indx,
cmt_seq,
concat(cmt_cmmt1,cmt_cmmt2,cmt_cmmt3,cmt_cmmt4,cmt_cmmt5,cmt_cmmt6,cmt_cmmt7,cmt_cmmt8,cmt_cmmt9,cmt_cmmt10,cmt_cmmt11,cmt_cmmt12,cmt_cmmt13,cmt_cmmt14,cmt_cmmt15) as cmt_cmmt,
cmt_lang
from
v_cmt_det_act
)
select
cmt_domain,
cmt_indx,
--FOR XML PATH returns a one record XML datatype in which all records are "unioned".
-- However, special characters are encoded as excape sequences, i.e. "<" becomes "<"
-- To reverse the encoding, get the entire root document/element as a value using the VALUE function.
-- The [1] is required because VALUE expects a singleton, i.e. exactly ONE element
rtrim(substring(
(
select char(10) + '[PAGE ' + cast(x2.cmt_seq as nvarchar(2)) + ']' + char(10) + x2.cmt_cmmt -- a field with no name gets no XML element to enclosed it. In short, the XML tag is removed
from x x2
where x1.cmt_domain = x2.cmt_domain and x1.cmt_indx = x2.cmt_indx
order by x2.cmt_seq
for XML PATH (''), type -- the brackets indicate the XML root element. Specifying an empty string as root element name removes it
).value('(/)[1]','nvarchar(max)')
,2,100000)) as cmt_cmmt
from
(
select distinct cmt_domain, cmt_indx
from x
) x1

Query Active Directory from SQL Server

Get active directory server and LDAP servers in a domain
nltest /dclist:sub.dom.com

results amongst others in:
\\dcserver1.sub.dom.com -> dcserver1 is the AD server

Analyse active directory structure
Use Sysinternals ADExplorer to analyse the structure of active directory. Connect to server: dcserver1

Users in a group from command prompt
net group adtestgroup /domain

Active Directory structure in FROM clause
E.g. FROM "LDAP://sub.dom.com/OU=Groups,OU=Global,DC=sub,DC=dom,DC=com"

LDAP = case-sensitive protocol name, always this value
Before the slash: sub.dom.com, is the domain for which to find the AD structure
After the slash: the part of the tree to search. Specified from right to left. So in the example, from parent to child:
com -> dom -> sub -> Global -> Groups -> ADTestGroup, searches the nodes beneath the lowest level, which is ADTestGroup

The names ("dom", "Groups", etc), type ("OU", "CN", "DC") and depth of the levels are dynamic. So check with ADExplorer how AD is configured in the specific instance.

Users in a group from SQL Server
First, add a linked server to ADSI (fill with correct password for DOMAIN\ACCOUNT):

exec sp_addlinkedserver @server = 'ADSI', @srvproduct = 'Active Directory Services 2.5', @provider = 'ADSDSOObject', @datasrc = 'adsdatasource'
exec sp_addlinkedsrvlogin @rmtsrvname = 'ADSI', @useself = 'False', @locallogin = null, @rmtuser = 'DOMAIN\ACCOUNT', @rmtpassword = '********'

Select users in group "adtestgroup":

select * from openquery(ADSI,'
    SELECT objectCategory, cn, sn, mail, name, department, company
    FROM ''LDAP://sub.dom.com/DC=sub,DC=dom,DC=com''
    WHERE MemberOf=''CN=adtestgroup,OU=Groups,OU=Global,DC=sub,DC=dom,DC=com''
    ORDER BY cn
')

Unfortunately, dynamic sql is not possible inside a udf, so a procedure is needed:

alter procedure cst_usersingroup(@grp nvarchar(100))
as
begin
    declare @sql as nvarchar(max) = '
    select samaccountname as accountname, name, mail, department, title, company
    from openquery(ADSI,''
        SELECT objectCategory, cn, sn, mail, name, department, company, title, samaccountname
        FROM ''''LDAP://sub.dom.com/DC=sub,DC=dom,DC=com''''
        WHERE MemberOf=''''CN=' + @grp + ',OU=Groups,OU=Global,DC=sub,DC=dom,DC=com'''''')
    x
    order by samaccountname'

    exec(@sql)
end

Select all BI_* and EDW_* groups from active directory:

select lower(cn) as grp, whencreated, whenchanged, distinguishedname
from openquery(ADSI,'
    SELECT cn, distinguishedname, whencreated, whenchanged
    FROM ''LDAP://sub.dom.com/OU=Groups,OU=Global,DC=sub,DC=dom,DC=com''
    where CN=''bi_*'' or  CN=''edw_*''
') x
order by cn

Notes:

• In most cases, all domain users can query AD using the basic search method specified before. More advanced AD search methods might be disabled and require special rights

Create schema with separate autorization

Perform the following steps in SQL Server to create a schema called "bi" and give create view rights to user "domain\testuser":

CREATE SCHEMA [bi]
CREATE ROLE [db_bischema]
ALTER AUTHORIZATION ON SCHEMA::bi TO [db_bischema]
GRANT CREATE VIEW TO [db_bischema] -- "create view" includes the power to drop views
ALTER ROLE [db_bischema] ADD MEMBER [domain\testuser]

Advantages:

• If views are removed from or added to the schema, autorization does not need to be added to each individual view
• Give certain users the rights to (re)create the views inside the schema, without the risk of modification of the rest of the database

Pass table-valued parameter from VB.Net

1) Create table type itemtabletype for a table with one, unique clustered column called "item":

create type itemtabletype
as table (
item nvarchar(30) not null,
primary key clustered (item)
)

2) Create a procedure cst_testtabletype that has itemtable type as a table-valued input parameter:

create procedure cst_testtabletype

( @tv itemtabletype readonly )

as

begin

declare @itemcount int = 0

select @itemcount = count(*) from @tv

select concat('succesfully parsed item ',item) from @tv

end

The procedure returns all items provided in the table-valued parameter prefixed with the words "successfully parsed item ".

3) Give execution rights to user(s) and/or groups that are allowed to execute the procedure and table type:

grant exec on type::itemtabletype to [User]
grant exec on cst_testtabletype to [User]

4) Finally, create the following test routine  in VB.Net:

    Public Sub Start()

        'create data table
        Dim Table1 As DataTable

        'create a table named tmptbl
        Table1 = New DataTable("tmptbl")

        Dim row As DataRow

        Try

            'declare a column named item
            Dim item As DataColumn = New DataColumn("item")

            'setting the datatype for the column
            item.DataType = System.Type.GetType("System.String")

            'adding the column to table
            Table1.Columns.Add(item)

            'declaring a new row
            Dim i As Integer
            For i = 1 To 1000
                row = Table1.NewRow()
                row.Item("item") = i
                Table1.Rows.Add(row)
            Next

        Catch

        End Try

        'establishing connection. you need to provide password for SQL server
        Dim myConnection = New SqlConnection()

        myConnection.ConnectionString = "Server=mySQLServer;Database=myDB;Trusted_Connection=True;"

        Try

            myConnection.Open()

            Dim myCommand As New SqlCommand("cst_testtabletype", myConnection)

            ' the table-valued parameter is called @tv for cst_testtabletype
            Dim myparam As SqlParameter = myCommand.Parameters.Add("@tv", SqlDbType.Structured)
            'create parameter
            myCommand.CommandType = CommandType.StoredProcedure
            myparam.Value = Table1

            Dim dr As SqlDataReader = myCommand.ExecuteReader()
            While dr.Read()
                Debug.Print(dr.Item(0))
            End While

        Catch ex As SqlException

            MsgBox(ex.Message)

        End Try

    End Sub

IsNumeric() stinks

Read the thoroughly insightful and funny post by Mike Teevee as a reaction to the technet documentation with regards to the useless ISNUMERIC() function introduced in SQL Server 2012:
http://technet.microsoft.com/en-us/library/ms186272.aspx

Select "SQL Server 2012" under "Other Versions", otherwise the post is hidden.

Silverlight not working because of font issue

If Silverlight is installed correctly, but still not working, check for Windows Fonts with an ampersand in the name and remove them:

http://connect.microsoft.com/VisualStudio/feedback/details/719317/silverlight-5-error-1001-font-filename-issue

Uneasy way to determine proxy server

Open a command prompt and perform the following steps:

1. Inspect the output of netstat -an | find "EST" (short from 'ESTABLISHED')
2. Go to a fresh site (one that you have not recently visited)
3. Run the netstat command again, looking for the new connection. It might look like:
   TCP 192.168.1.1:1989 192.168.1.88:8080 ESTABLISHED

(credits go to Royce Williams)

Alternatively, install TCPView and look at the remote address and remote port columns for the browser connections.