crazyman
asked on
String concat order problem when grouping
Hi
I have written a CLR string concat function, the problem is when using it in my sql statement the order of the items that are concatenated seems to be dependant on the column itself.
How to be explicit on the order ?
My current sql
SELECT cast(dbo.StringConcat(cast (descripti on as varchar)) as varchar(max)) AS [description],
[vw_ECS_OM_JobNotes_Job].[ reference] AS [reference] FROM [vw_ECS_OM_JobNotes_Job]
GROUP BY [vw_ECS_OM_JobNotes_Job].[ reference]
results in the actual description being alphabetically appended to the aggregate result, however i have another column called sequence that i need to be the order...
I tried ordering inside the view but this doesnt work either ?
I have written a CLR string concat function, the problem is when using it in my sql statement the order of the items that are concatenated seems to be dependant on the column itself.
How to be explicit on the order ?
My current sql
SELECT cast(dbo.StringConcat(cast
[vw_ECS_OM_JobNotes_Job].[
GROUP BY [vw_ECS_OM_JobNotes_Job].[
results in the actual description being alphabetically appended to the aggregate result, however i have another column called sequence that i need to be the order...
I tried ordering inside the view but this doesnt work either ?
I've never done a CLR function, but my initial guess is that the result set, including the concatenation is all done in the first pass and then the results are sorted. By then, of course, it's too late. If that's the case, you may need to select your "driving data set" into a temp table (ordered) and then select off that, including the concat column.
ASKER
Ive debugged the clr function and it clearly adds the items in what seems to be a random order...
Ive attached my function, even when concatinsating on an id column it sometimes does it in a random order...
Any better contactinate aggregate functions around, ive seen some but they almost all require the function to have knowledge of the calling table, i need something generic...
Ive attached my function, even when concatinsating on an id column it sometimes does it in a random order...
Any better contactinate aggregate functions around, ive seen some but they almost all require the function to have knowledge of the calling table, i need something generic...
using Microsoft.SqlServer.Server;
using System.Data.SqlTypes;
using System.IO;
using System.Diagnostics;
namespace Peverel.SqlExtensions
{
[SqlUserDefinedAggregate(
//use CLR serialization to serialize the intermediate result.
Format.UserDefined,
//Optimizer property:
IsInvariantToNulls = true,
//Optimizer property:
IsInvariantToDuplicates = false,
//Optimizer property:
IsInvariantToOrder = false,
//Maximum size in bytes of persisted value:
MaxByteSize = 8000)]
public class StringConcat : IBinarySerialize
{
/// <summary>
/// Variable holds intermediate result of the concatenation
/// </summary>
private StringBuilder intermediateResult;
/// <summary>
/// Initialize the internal data structures
/// </summary>
public void Init()
{
intermediateResult = new StringBuilder();
}
/// <summary>
/// Accumulate the next value, nop if the value is null
/// </summary>
/// <param name="value"></param>
public void Accumulate(SqlString value)
{
if (value.IsNull)
{
return;
}
intermediateResult.Append(value.Value).Append(' ');
}
/// <summary>
/// Merge the partially computed aggregate with this aggregate.
/// </summary>
/// <param name="other"></param>
public void Merge(StringConcat other)
{
intermediateResult.Append(other.intermediateResult);
}
/// <summary>
/// Called at end of aggregation, to return results.
/// </summary>
/// <returns></returns>
public SqlString Terminate()
{
string output = string.Empty;
//Delete the trailing comma, if any .
if (intermediateResult != null && intermediateResult.Length > 0)
output = intermediateResult.ToString(0, intermediateResult.Length - 1);
return new SqlString(output);
}
public void Read(BinaryReader r)
{
intermediateResult = new StringBuilder(r.ReadString());
}
public void Write(BinaryWriter w)
{
w.Write(intermediateResult.ToString());
}
}
}
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
Hmmm
I had an idea, if i can sort the results in my clr object then i can control the output, however you cant pass the sort column data in as another param, so i thought if you appended it with a delimiter at the start of the data and used it to sort the result.
Anyway it is a bit of a bodge job but it works great...
Sql looks like
SELECT cast(dbo.StringConcat(cast (sequence as varchar) + '|' + cast(description as varchar)) as varchar(max)) AS [description],
[vw_ECS_OM_JobNotes_Job].[ reference] AS [reference] FROM [vw_ECS_OM_JobNotes_Job]
GROUP BY [vw_ECS_OM_JobNotes_Job].[ reference]
Code attached..
I had an idea, if i can sort the results in my clr object then i can control the output, however you cant pass the sort column data in as another param, so i thought if you appended it with a delimiter at the start of the data and used it to sort the result.
Anyway it is a bit of a bodge job but it works great...
Sql looks like
SELECT cast(dbo.StringConcat(cast
[vw_ECS_OM_JobNotes_Job].[
GROUP BY [vw_ECS_OM_JobNotes_Job].[
Code attached..
using System;
using System.Collections.Generic;
using System.Collections;
using System.Text;
using Microsoft.SqlServer.Server;
using System.Data.SqlTypes;
using System.IO;
using System.Diagnostics;
namespace Peverel.SqlExtensions
{
[SqlUserDefinedAggregate(
//use CLR serialization to serialize the intermediate result.
Format.UserDefined,
//Optimizer property:
IsInvariantToNulls = true,
//Optimizer property:
IsInvariantToDuplicates = false,
//Optimizer property:
IsInvariantToOrder = false,
//Maximum size in bytes of persisted value:
MaxByteSize = 8000)]
public class StringConcat : IBinarySerialize
{
private class AggregateData
{
public string Value;
public string RawValue;
public int Order = 0;
private AggregateData() { }
public static AggregateData Parse(string value)
{
int order = 0;
string[] parts = value.Split('|');
AggregateData ad = new AggregateData();
ad.Value = value;
ad.RawValue = value;
if (parts.Length > 1 && int.TryParse(parts[0],out order))
{
ad.Order = order;
ad.Value = value.Substring(order.ToString().Length+1);
}
return ad;
}
}
private List<AggregateData> values;
public void Init()
{
this.values = new List<AggregateData>();
}
public void Accumulate(SqlString value)
{
this.values.Add(AggregateData.Parse(value.Value));
}
public void Merge(StringConcat value)
{
this.values.AddRange(value.values);
}
public SqlString Terminate()
{
this.values.Sort(new Comparison<AggregateData>(SortValue));
StringBuilder sb = new StringBuilder();
foreach (AggregateData ad in this.values)
{
//For debug you can see the order val
sb.Append(String.Format("{0}|{1}",ad.Order,ad.Value));
sb.Append(ad.Value);
}
return new SqlString(sb.ToString());
}
private int SortValue(AggregateData a, AggregateData b)
{
return a.Order.CompareTo(b.Order);
}
public void Read(BinaryReader r)
{
int itemCount = r.ReadInt32();
this.values = new List<AggregateData>(itemCount);
for (int i = 0; i <= itemCount - 1; i++)
{
this.values.Add(AggregateData.Parse(r.ReadString()));
}
}
public void Write(BinaryWriter w)
{
w.Write(this.values.Count);
foreach (AggregateData ad in this.values)
{
w.Write(ad.RawValue);
}
}
}
}
ASKER
oops just incase anyone else uses this solution comment out the line
sb.Append(String.Format("{ 0}|{1}",ad .Order,ad. Value));
sb.Append(String.Format("{
ASKER
SELECT cast(dbo.StringConcat(cast
[reference] AS [reference] FROM
(
select top 100 percent * from
[vw_ECS_OM_JobNotes_Job]
order by reference,sequence) as sub
GROUP BY [reference]