How can I make this more efficient? (combining DataSet results with the results of a DB lookup.)

Discussion in 'ASP .Net' started by Ken Fine, Jul 18, 2008.

  1. Ken Fine

    Ken Fine Guest

    This is a question that someone familiar with ASP.NET and ADO.NET DataSets
    and DataTables should be able to answer fairly easily. The basic question is
    how I can efficiently match data from one dataset to data in a second
    dataset, using a common key. I will first describe the problem in words and
    then I will show my code, which has most of the solution done already.

    I have built an ASP.NET that queries an Index Server and returns a DataSet
    corresponding to a bunch of PDF files. I bind this DataSet to an ASP.NET
    gridview, but I want to include more information than is in Index Server
    alone (and no, I don't want to use IS custom properties, nor do I want to
    use MSFT's great new free search server quite yet.) . I want to do a second
    database lookup to gather more information and I have a good idea how to do
    that. I have successfully manually added extra columns to my DataSet's
    datatable and populated them with database data in an inefficient way which
    I will describe in the next paragraph.

    The filenames associated with the PDFs in the Index Server index include
    database Primary Keys and my application parses those keys out. I use those
    primary keys to subsequently do a second collection of database lookups.
    Currently just for test purposes I have the application doing dozens of
    lookups: it's querying the database on each iteration of a loop. This sucks
    from a performance standpoint. In anticipation of doing a single query of
    all necessary lookups, I have made it so that the system builds a functional
    SQL query string that pulls all necessary records with additional metadata.

    What I need to know is:
    1) Syntax for creating a new dataset based on my SQL string. I normally use
    the EntitySpaces ORM framework so my experience with Datasets is very weak.
    2) How to match the parts of the new dataset into my existing dataset in an
    EFFICIENT way. My code below will show a successful solution based on a very
    INEFFICIENT way, using repeated lookups. I expect the solution will involve
    repeated filterings of the single dataset based on the primary key.

    Now to my existing, inefficient code. I will walk you through it.

    // Filling a DataSet called results, and manually adding some additional
    columns to it:

    if (dbAdapter != null) dbAdapter.SelectCommand.CommandText =
    GetUwitmPdfArchiveQuery;
    DataSet ds = new DataSet("Results");
    dbAdapter.Fill(ds);
    DataColumn dcolColumn = new DataColumn("ChrisSummary",
    typeof(string));
    ds.Tables[0].Columns.Add(dcolColumn);
    DataColumn dcolColumn2 = new DataColumn("ChrisTitle",
    typeof(string));
    ds.Tables[0].Columns.Add(dcolColumn2);
    DataColumn dcolColumn3 = new DataColumn("PublicationTitle",
    typeof(string));
    ds.Tables[0].Columns.Add(dcolColumn3);
    DataColumn dcolColumn4 = new DataColumn("OriginalArticleUrl",
    typeof(string));
    ds.Tables[0].Columns.Add(dcolColumn4);

    // [empty variables are initialized]

    // start looking through the datarows. There is a DB primary key in the data
    one of the columns of the datarow . I grab that key and assign it the name
    tempInt

    foreach (DataRow dr in ds.Tables[0].Rows)
    {
    int tempInt =
    Convert.ToInt32(dr[4].ToString().Replace(".pdf", ""));
    try
    {

    // I start concatenating a SQL string of all of the primary keys
    listOfIds = listOfIds + " OR ContentID ='" +
    dr[4].ToString().Replace(".pdf", "") + "'";

    // This is the inefficient part. I'm using that key along with my
    EntitySpaces ORM framework to do lookups each time this loop is iterated,
    and populate my manually added rows. I want instead to do a single lookup
    and use a second filtered DataSet to populate my manually added rows with DB
    information. So I'm going to have to loop through once to get enough info to
    build my SQL string, then do something else to build the DataSet I need.
    This is where I need the help.

    Contentitems ci = new Contentitems();
    ci.LoadByPrimaryKey(Convert.ToInt32(tempInt));
    dr["OriginalArticleUrl"] = ci.ConUrl;
    dr["ChrisSummary"] = ci.ConSummary;
    dr["ChrisTitle"] = ci.ConTitle;
    dr["PublicationTitle"] = ci.ConSource;

    // I finish my SQL statement and whack off some extraneous stuff from my SQL
    string

    listOfIds = listOfIds.Substring(0, listOfIds.Length - 25);

    string sqlQueryIds = "Select * From Contentitems WHERE
    ContentID='99999999' " + listOfIds;


    Would very much appreciate help making the above code efficient! Thanks!!!

    -KF
     
    Ken Fine, Jul 18, 2008
    #1
    1. Advertising

  2. The easiest/fastest way is to use a filter on the datatable. I'll qualify
    this by saying I haven't tested this against LINQ yet.

    Roughly, you will pass in a sql like query to the datatable's select method
    and it will return the rows that match the filter. Here is a complete
    example from MSDN.

    private void GetRowsByFilter()
    {
    DataTable table = DataSet1.Tables["Orders"];
    // Presuming the DataTable has a column named Date.
    string expression;
    expression = "Date > #1/1/00#";
    DataRow[] foundRows;

    // Use the Select method to find all rows matching the filter.
    foundRows = table.Select(expression);

    // Print column 0 of each returned row.
    for(int i = 0; i < foundRows.Length; i ++)
    {
    Console.WriteLine(foundRows[0]);
    }
    }

    --

    Regards,
    Alvin Bruney [MVP ASP.NET]

    [Shameless Author plug]
    Download OWC Black Book, 2nd Edition
    Exclusively on www.lulu.com/owc $15.00
    Need a free copy of VSTS 2008 w/ MSDN Premium?
    http://msmvps.com/blogs/alvin/Default.aspx
    -------------------------------------------------------


    "Ken Fine" <> wrote in message
    news:...
    > This is a question that someone familiar with ASP.NET and ADO.NET DataSets
    > and DataTables should be able to answer fairly easily. The basic question
    > is how I can efficiently match data from one dataset to data in a second
    > dataset, using a common key. I will first describe the problem in words
    > and then I will show my code, which has most of the solution done already.
    >
    > I have built an ASP.NET that queries an Index Server and returns a DataSet
    > corresponding to a bunch of PDF files. I bind this DataSet to an ASP.NET
    > gridview, but I want to include more information than is in Index Server
    > alone (and no, I don't want to use IS custom properties, nor do I want to
    > use MSFT's great new free search server quite yet.) . I want to do a
    > second database lookup to gather more information and I have a good idea
    > how to do that. I have successfully manually added extra columns to my
    > DataSet's datatable and populated them with database data in an
    > inefficient way which I will describe in the next paragraph.
    >
    > The filenames associated with the PDFs in the Index Server index include
    > database Primary Keys and my application parses those keys out. I use
    > those primary keys to subsequently do a second collection of database
    > lookups. Currently just for test purposes I have the application doing
    > dozens of lookups: it's querying the database on each iteration of a
    > loop. This sucks from a performance standpoint. In anticipation of doing a
    > single query of all necessary lookups, I have made it so that the system
    > builds a functional SQL query string that pulls all necessary records with
    > additional metadata.
    >
    > What I need to know is:
    > 1) Syntax for creating a new dataset based on my SQL string. I normally
    > use the EntitySpaces ORM framework so my experience with Datasets is very
    > weak.
    > 2) How to match the parts of the new dataset into my existing dataset in
    > an EFFICIENT way. My code below will show a successful solution based on a
    > very INEFFICIENT way, using repeated lookups. I expect the solution will
    > involve repeated filterings of the single dataset based on the primary
    > key.
    >
    > Now to my existing, inefficient code. I will walk you through it.
    >
    > // Filling a DataSet called results, and manually adding some additional
    > columns to it:
    >
    > if (dbAdapter != null) dbAdapter.SelectCommand.CommandText =
    > GetUwitmPdfArchiveQuery;
    > DataSet ds = new DataSet("Results");
    > dbAdapter.Fill(ds);
    > DataColumn dcolColumn = new DataColumn("ChrisSummary",
    > typeof(string));
    > ds.Tables[0].Columns.Add(dcolColumn);
    > DataColumn dcolColumn2 = new DataColumn("ChrisTitle",
    > typeof(string));
    > ds.Tables[0].Columns.Add(dcolColumn2);
    > DataColumn dcolColumn3 = new DataColumn("PublicationTitle",
    > typeof(string));
    > ds.Tables[0].Columns.Add(dcolColumn3);
    > DataColumn dcolColumn4 = new DataColumn("OriginalArticleUrl",
    > typeof(string));
    > ds.Tables[0].Columns.Add(dcolColumn4);
    >
    > // [empty variables are initialized]
    >
    > // start looking through the datarows. There is a DB primary key in the
    > data one of the columns of the datarow . I grab that key and assign it the
    > name tempInt
    >
    > foreach (DataRow dr in ds.Tables[0].Rows)
    > {
    > int tempInt =
    > Convert.ToInt32(dr[4].ToString().Replace(".pdf", ""));
    > try
    > {
    >
    > // I start concatenating a SQL string of all of the primary keys
    > listOfIds = listOfIds + " OR ContentID ='" +
    > dr[4].ToString().Replace(".pdf", "") + "'";
    >
    > // This is the inefficient part. I'm using that key along with my
    > EntitySpaces ORM framework to do lookups each time this loop is iterated,
    > and populate my manually added rows. I want instead to do a single lookup
    > and use a second filtered DataSet to populate my manually added rows with
    > DB information. So I'm going to have to loop through once to get enough
    > info to build my SQL string, then do something else to build the DataSet I
    > need. This is where I need the help.
    >
    > Contentitems ci = new Contentitems();
    > ci.LoadByPrimaryKey(Convert.ToInt32(tempInt));
    > dr["OriginalArticleUrl"] = ci.ConUrl;
    > dr["ChrisSummary"] = ci.ConSummary;
    > dr["ChrisTitle"] = ci.ConTitle;
    > dr["PublicationTitle"] = ci.ConSource;
    >
    > // I finish my SQL statement and whack off some extraneous stuff from my
    > SQL string
    >
    > listOfIds = listOfIds.Substring(0, listOfIds.Length - 25);
    >
    > string sqlQueryIds = "Select * From Contentitems WHERE
    > ContentID='99999999' " + listOfIds;
    >
    >
    > Would very much appreciate help making the above code efficient! Thanks!!!
    >
    > -KF
    >
    >
    >
     
    Alvin Bruney [ASP.NET MVP], Jul 19, 2008
    #2
    1. Advertising

  3. Hi KF,

    Based on your description, I understand that you have a DataSet/DataTable
    generqted via some query from index server. And since each of the record
    may have some additional data/fields in database, currently you're
    performing a data access query from database server when looping each of
    the DataSet/DataTable record, correct?

    Yes, I can see that there does be some performance overhead when the record
    count is quite large. In my opinion, I think you can consider the following
    approaches to impove the query performance(both of them require more memory
    space since you'll need to cache those additional fields locally in memory).

    1. Before you loop the first Dataset/DataTable(generated from index server
    ), you can perform another query against the database which return all the
    necessary records(contain the required additional fields), you can sort
    them via the primary key(you want to search against later), this sort is at
    T-SQL level. You can also sort it in DataTable , but I think T-SQL layer
    will be more efficient:

    #DataTable..::.Select Method
    http://msdn.microsoft.com/en-us/library/system.data.datatable.select.aspx

    #Filtering and Sorting with the DataTable Select Me
    http://www.developerfusion.co.uk/show/4703/2/

    After that , you get a sorted record sequence in memory (based primarykey).
    Later in your code which loop each record in the index server records, you
    can perform a binary search like approach in the returned records set above.

    2. If the data in database is quite simple(not many fields), you can even
    directly store all the fields into hashtable based on the certain primary
    key. Later you can directly access HashTable for the requied data. This is
    extremely quick and efficient.

    Sincerely,

    Steven Cheng

    Microsoft MSDN Online Support Lead


    Delighting our customers is our #1 priority. We welcome your comments and
    suggestions about how we can improve the support we provide to you. Please
    feel free to let my manager know what you think of the level of service
    provided. You can send feedback directly to my manager at:
    .

    ==================================================
    Get notification to my posts through email? Please refer to
    http://msdn.microsoft.com/subscriptions/managednewsgroups/default.aspx#notif
    ications.

    ==================================================
    This posting is provided "AS IS" with no warranties, and confers no rights.
    --------------------
    >From: "Ken Fine" <>
    >Subject: How can I make this more efficient? (combining DataSet results

    with the results of a DB lookup.)
    >Date: Fri, 18 Jul 2008 15:25:18 -0700


    >
    >This is a question that someone familiar with ASP.NET and ADO.NET DataSets
    >and DataTables should be able to answer fairly easily. The basic question

    is
    >how I can efficiently match data from one dataset to data in a second
    >dataset, using a common key. I will first describe the problem in words

    and
    >then I will show my code, which has most of the solution done already.
    >
    >I have built an ASP.NET that queries an Index Server and returns a DataSet
    >corresponding to a bunch of PDF files. I bind this DataSet to an ASP.NET
    >gridview, but I want to include more information than is in Index Server
    >alone (and no, I don't want to use IS custom properties, nor do I want to
    >use MSFT's great new free search server quite yet.) . I want to do a

    second
    >database lookup to gather more information and I have a good idea how to

    do
    >that. I have successfully manually added extra columns to my DataSet's
    >datatable and populated them with database data in an inefficient way

    which
    >I will describe in the next paragraph.
    >
    >The filenames associated with the PDFs in the Index Server index include
    >database Primary Keys and my application parses those keys out. I use

    those
    >primary keys to subsequently do a second collection of database lookups.
    >Currently just for test purposes I have the application doing dozens of
    >lookups: it's querying the database on each iteration of a loop. This

    sucks
    >from a performance standpoint. In anticipation of doing a single query of
    >all necessary lookups, I have made it so that the system builds a

    functional
    >SQL query string that pulls all necessary records with additional metadata.
    >
    >What I need to know is:
    >1) Syntax for creating a new dataset based on my SQL string. I normally

    use
    >the EntitySpaces ORM framework so my experience with Datasets is very weak.
    >2) How to match the parts of the new dataset into my existing dataset in

    an
    >EFFICIENT way. My code below will show a successful solution based on a

    very
    >INEFFICIENT way, using repeated lookups. I expect the solution will

    involve
    >repeated filterings of the single dataset based on the primary key.
    >
    >Now to my existing, inefficient code. I will walk you through it.
    >
    >// Filling a DataSet called results, and manually adding some additional
    >columns to it:
    >
    > if (dbAdapter != null) dbAdapter.SelectCommand.CommandText =
    >GetUwitmPdfArchiveQuery;
    > DataSet ds = new DataSet("Results");
    > dbAdapter.Fill(ds);
    > DataColumn dcolColumn = new DataColumn("ChrisSummary",
    >typeof(string));
    > ds.Tables[0].Columns.Add(dcolColumn);
    > DataColumn dcolColumn2 = new DataColumn("ChrisTitle",
    >typeof(string));
    > ds.Tables[0].Columns.Add(dcolColumn2);
    > DataColumn dcolColumn3 = new DataColumn("PublicationTitle",
    >typeof(string));
    > ds.Tables[0].Columns.Add(dcolColumn3);
    > DataColumn dcolColumn4 = new DataColumn("OriginalArticleUrl",
    >typeof(string));
    > ds.Tables[0].Columns.Add(dcolColumn4);
    >
    >// [empty variables are initialized]
    >
    >// start looking through the datarows. There is a DB primary key in the

    data
    >one of the columns of the datarow . I grab that key and assign it the name
    >tempInt
    >
    > foreach (DataRow dr in ds.Tables[0].Rows)
    > {
    > int tempInt =
    >Convert.ToInt32(dr[4].ToString().Replace(".pdf", ""));
    > try
    > {
    >
    >// I start concatenating a SQL string of all of the primary keys
    > listOfIds = listOfIds + " OR ContentID ='" +
    >dr[4].ToString().Replace(".pdf", "") + "'";
    >
    >// This is the inefficient part. I'm using that key along with my
    >EntitySpaces ORM framework to do lookups each time this loop is iterated,
    >and populate my manually added rows. I want instead to do a single lookup
    >and use a second filtered DataSet to populate my manually added rows with

    DB
    >information. So I'm going to have to loop through once to get enough info

    to
    >build my SQL string, then do something else to build the DataSet I need.
    >This is where I need the help.
    >
    > Contentitems ci = new Contentitems();
    > ci.LoadByPrimaryKey(Convert.ToInt32(tempInt));
    > dr["OriginalArticleUrl"] = ci.ConUrl;
    > dr["ChrisSummary"] = ci.ConSummary;
    > dr["ChrisTitle"] = ci.ConTitle;
    > dr["PublicationTitle"] = ci.ConSource;
    >
    >// I finish my SQL statement and whack off some extraneous stuff from my

    SQL
    >string
    >
    > listOfIds = listOfIds.Substring(0, listOfIds.Length - 25);
    >
    > string sqlQueryIds = "Select * From Contentitems WHERE
    >ContentID='99999999' " + listOfIds;
    >
    >
    >Would very much appreciate help making the above code efficient! Thanks!!!
    >
    >-KF
    >
    >
    >
    >
     
    Steven Cheng [MSFT], Jul 21, 2008
    #3
  4. Hi KF,

    Have you got any further ideas on this or have you made the decision which
    approach to use? If there is anything else we can help, please feel free to
    post here.

    Sincerely,

    Steven Cheng

    Microsoft MSDN Online Support Lead

    Delighting our customers is our #1 priority. We welcome your comments and
    suggestions about how we can improve the support we provide to you. Please
    feel free to let my manager know what you think of the level of service
    provided. You can send feedback directly to my manager at:
    .

    ==================================================
    Get notification to my posts through email? Please refer to
    http://msdn.microsoft.com/subscriptions/managednewsgroups/default.aspx#notif
    ications.
    ==================================================
    This posting is provided "AS IS" with no warranties, and confers no rights.
    --------------------

    >From: (Steven Cheng [MSFT])
    >Organization: Microsoft
    >Date: Mon, 21 Jul 2008 03:54:01 GMT
    >Subject: RE: How can I make this more efficient? (combining DataSet

    results with the results of a DB lookup.)

    >
    >Hi KF,
    >
    >Based on your description, I understand that you have a DataSet/DataTable
    >generqted via some query from index server. And since each of the record
    >may have some additional data/fields in database, currently you're
    >performing a data access query from database server when looping each of
    >the DataSet/DataTable record, correct?
    >
    >Yes, I can see that there does be some performance overhead when the

    record
    >count is quite large. In my opinion, I think you can consider the

    following
    >approaches to impove the query performance(both of them require more

    memory
    >space since you'll need to cache those additional fields locally in

    memory).
    >
    >1. Before you loop the first Dataset/DataTable(generated from index server
    >), you can perform another query against the database which return all the
    >necessary records(contain the required additional fields), you can sort
    >them via the primary key(you want to search against later), this sort is

    at
    >T-SQL level. You can also sort it in DataTable , but I think T-SQL layer
    >will be more efficient:
    >
    >#DataTable..::.Select Method
    >http://msdn.microsoft.com/en-us/library/system.data.datatable.select.aspx
    >
    >#Filtering and Sorting with the DataTable Select Me
    >http://www.developerfusion.co.uk/show/4703/2/
    >
    >After that , you get a sorted record sequence in memory (based

    primarykey).
    >Later in your code which loop each record in the index server records, you
    >can perform a binary search like approach in the returned records set

    above.
    >
    >2. If the data in database is quite simple(not many fields), you can even
    >directly store all the fields into hashtable based on the certain primary
    >key. Later you can directly access HashTable for the requied data. This is
    >extremely quick and efficient.
    >
    >Sincerely,
    >
    >Steve>>
    >
    >
     
    Steven Cheng [MSFT], Jul 23, 2008
    #4
    1. Advertising

Want to reply to this thread or ask your own question?

It takes just 2 minutes to sign up (and it's free!). Just click the sign up button to choose a username and then you can ask your own questions on the forum.
Similar Threads
  1. Roy
    Replies:
    6
    Views:
    403
    Scott Allen
    Feb 16, 2005
  2. Replies:
    12
    Views:
    538
    Daniel T.
    Dec 4, 2007
  3. Bill David
    Replies:
    70
    Views:
    1,709
    James Kanze
    Sep 25, 2008
  4. PerlFAQ Server
    Replies:
    0
    Views:
    112
    PerlFAQ Server
    Jan 28, 2011
  5. PerlFAQ Server
    Replies:
    0
    Views:
    109
    PerlFAQ Server
    Feb 22, 2011
Loading...

Share This Page