angeshwar (1/13/2014)

---

Hi,

Yes, corrected the index for ENGINE_ENTITIES

CREATE TABLE [dbo].[ENGINE_ENTITIES](

[ID] [bigint] IDENTITY(1,1) NOT NULL,

[CREATED_DATE] [datetime] NULL,

[ENTITY_ID] [varchar](255) NULL,

[ENTITY_TYPE] [varchar](255) NULL,

[PROCESS_INSTANCE_ID] [bigint] NULL,

[STRING_DATA] [varchar](255) NULL,

[UPDATED_DATE] [datetime] NULL,

[XML_DATA] [text] NULL,

PRIMARY KEY NONCLUSTERED

(

[ID]...

angeshwar (1/8/2014)

---

Structure of ENGINE_ENTITIES

CREATE TABLE [dbo].[ENGINE_ENTITIES](

[ID] [bigint] IDENTITY(1,1) NOT NULL,

[CREATED_DATE] [datetime] NULL,

[ENTITY_ID] [varchar](255) NULL,

[ENTITY_TYPE] [varchar](255) NULL,

[PROCESS_INSTANCE_ID] [bigint] NULL,

[STRING_DATA] [varchar](255) NULL,

[UPDATED_DATE] [datetime] NULL,

[XML_DATA] [text] NULL,

PRIMARY KEY CLUSTERED

(

[ID] ASC

)WITH (PAD_INDEX ...

Hi. Thank you for the effort.

My query will return queries from cached plans. I think you missed one point. You need to click query_plan column on each row to visualize...

Could you run this query to get cached queries for ENGINE_ENTITIES table.

This will return cached query plan which you can open by clicking it. After that save them in .sqlplan...

We are using hibernate and jpa with glassfish with no problems after adding this property to connection string. Varchars are sent as varchars. Previously they were sent as nvarchar.

Could you...

You were quicker than me 🙂

Without seeing actual or estimated execution plan my guess is that nonclustered index beginning with NODE_INSTANCE_ID was not exclusive enough to do index seek or...

angeshwar (1/6/2014)

---

I have updated my connection string to

jdbc.url=jdbc:jtds:sqlserver://XXXXX:1433/XXXXX;instance=MSSQLServer;sendStringParametersAsUnicode=false

I am still facing the deadlock.

Can confirm that 2 queries causing deadlock are #select#.

I am not able to find the actual values...

<inputbuf>

(@P0 nvarchar(4000),@P1 bigint)select nodeinstan0_.id as id35_, nodeinstan0_.BPMN_ELEMENT_ID...

I can see from info that you have provided that your queries will not use indexes. Your table has varchar fields and query is called with nvarchar.

You need to alter...

Couple of things.

Can't you just use constraints?

Your posted code seems to assume that there will be just one row coming to trigger, which might not be the case.

By how many of those hundred columns client wants to order? If number is limited then adding clustered primary key and index to each column helps.

Is there lot of changes...

What are you trying to do?

There must be easier way to this.

ALTER TRIGGER trgFatalToxicity ON Symptom

AFTER INSERT

SET NOCOUNT ON;

IF EXISTS (SELECT *

FROM inserted

WHERE inserted.Grade=5)

BEGIN

PRINT 'Patient died!,...

Do you have any kind of problems?

If not, I can't see any reason for trying to solve this.

I think you need to go trough items in list and check selected for each one of them. Generate SQL from that and use in (selected items separated with comma)...