#Indexing

Hi, folks!

Suppose you have a Caché class with %String property which contains relatively large text (from 10 to 2000 symbols).

The class:

Class Test.Duplicates Extends %Persistent 

{

Property Text As %String (MAXLEN = 2000);

}

And you have thousands of entries.

What are the best options to find entries which are duplicates on this property?

Hi,

I have a class with around 400k lines and 60 columns. Class storage is Cache SQL storage (Mapped from a global).

 I want to create multiple indices on certain fields.

I am familiar with two approaches:

1. Create a new map (Index type) on a pointer global.

2. Create a bitmap index

Which approach is more recommended to be used in the case I described? If there are any other approaches, I will be happy to hear.

Thanks :)

Hi!

I'd like to know if there are any issues if an index is inserted into a table without running the %BuildIndices() method.

It's important to note that data inserted before the index is not important for retrieval, so it's not a problem data inserted before the index don't show up in queries.

The reason why I'm asking this is that I'd like to avoid index reconstruction on big tables which I need to inser such index.

I'm using Cache 2018.1.

Thanks,

José

Hello,

I have a class with a "Unique" index (pxfactidIndex) on a %Numeric property (pxfactid) (partially-edit code snippet below):

PropertyAs%Library.Numeric

I see the following odd behavior when counting how many records have that property set to null:

• The first query (17) returns the number of records in the table. The 2nd query (18) returns what I expect: all records have non-null values for that "pxfactid" property. However, when I run the 3rd query (19) which counts all records with non-null values in that field, I get a much smaller, unexpected, number.

Hi, Community!

Consider you move data from one server to another or make a deployment with persistent data. What do you do with index globals?

Is it always better to rebuild them or there are some cases when it worth to take them too?

Is there a way to get a good performing index on a date field? I have tried various date property indexes and the query plan is always in a pretty high range. Below are query plan result values I have observed:

   StartDate > '2019-12-01'  --cost = 699168
   StartDate = '2019-12-21'  --cost 70666
   StartDate between '2019-12-21' and '2019-21-28' --cost = 492058

The query plans above were for type %TimeStamp.

Hi

1. Is it possible do define an index like that :

create index UIX on MyTable (Column1) where Column1 is not null

2. What happens if we add an index on a property that is NOT required, meanning that not all records will be indexed because we do not allow null subscripts ?

I'm facing a specific sort problem.
There are several thousands of articles sold all over. 
Users expect to get a description in local language sorted by their specific collation.

I'm using Cache SQL and want the ability to choose a specific index.

I've boiled the problem down to one table and simplified the query down to

SELECT *
FROM Registration.PatResp
WHERE SchedApptNum=8450022

SchedApptNum is indexed, but instead of using that column, "Show Plan" indicates that it's looping through the entire Registration.PatResp table on Id (the primary key for the table).

I've done a tune-table with no change.

SQL provides %IGNOREINDEX, which allows me to ignore a specific index, but I don't need that.  I need the reverse - %CHOOSEINDEX so that I can force SQL to use a specific index.

Hi!

Consider I have a class Package.Data with Property UniqueStringValue as %String.

I introduced the Index for this property:

 Index ValueIndex on UniqueStringValue [Unique];

It works well.  But if I try to check if there is an object with the certain value in code like this:

if ##class(Package.Data).ValueIndexExists(value)  

this expression fails, if value="value", even if there is an instance with instance.UniqueStingValue="Value"

How can I set the index to prevent saving case sensitive values in this class?

Hello. I need to transform a message

FROM:

MSH|^~\&|
SCH||61490||
PID|1||
RGS|1||1
AIS|1||
AIS|2||
AIS|3||
AIL|1||
AIP|1||

TO:

MSH|^~\&|
SCH||61490||
PID|1||
RGS|1||1
AIS|1||
AIL|1||
AIP|1||
RGS|1||1
AIS|2||
AIL|1||
AIP|1||
RGS|1||1
AIS|3||
AIL|1||
AIP|1||

The RGS, AIS, AIL and AIP are all under the RGS group. The one RGS segment that comes in will be copied across the group. If 3 AIS segments come in then I need 3 RGS groups, if 2 I need 2 RGS groups etc.

In my DTL (screenshot below) I have currently hardcoded the RGS index (1,2,3) but this will not be sufficient incase 4 AIS segments are sent in.

I want to store data in an index global without defining an index in an inherited class.
Example:

Class Aclass

Class AAclass Extends (%Persistent, ClassType.SuperClass, Aclass

Is there a way to inherit indexes? (without defining these indices) ^AAclassI("TXSSI" ...
Do you have any advice on how to do this?
Thank you

Hi,

I am new to Cache' MV but have extensive experience with other Pick flavors especially Unidata. 

I need to determine the impact of adding several indexes to a large file with over 51,000,000 records. 

On other systems, I could use FILE.STAT, ANALYZE.FILE and shell to the OS to determine how large the index file was. 

None of those seem to be available in Cache' MV. Shelling to the OS just tells me the size of CACHE.DAT.

What is the best way to determine what the disk impact would be if I added an index (CREATE-INDEX) to a file?

TIA,

Steve

Hi,
just want to know if indexes in IRIS / SQL are autobuilding.

I define an index in a table and then i costantly do a lot of insert/delete in that table. Will i face index problem? Do i need to rebuild it often?

Thanks

Hi guys!

Unique, PrimaryKey and IDKey?
In what contexts does it apply?

IDKey sets the registry key access to the store.
PrimaryKey, Unique, and IDKey define the uniqueness in the records, but what is correct?

I use everyone? What is the context of each?

Just wondering an Insight in the difference between these two indexes

IdKey / PrimaryKey
================= 

Property Identifier As %Integer

Index Index1 on Identifier [Idkey]

Index Index2 on Identifier [PrimaryKey]

What's the difference?

1. If I don't have Index1 and only have Index2,  then cache does still make its own id.
So how and why  do I ever use the PrimaryKey.  In Joins ??

Table1.Identifier = Table2.Identifier instead of Table1.Id = Table2.id ??
But I can still use Table1.Id = Table2.Id as cache still made one ID field

So where is PrimaryKey useful in cache?

I have a %Persistent class with properties that are of %SerialObject. I want to add an index to a property of the %SerialObject class.

Is this possible?

Cache version: Cache for Windows (x86-64) 2017.2.1 (Build 801_3U)

Good Afternoon,
I have a co-worker who is trying to run the below query via ODBC. The issue is that the query appears to be running extremely slow (nearly 2 hours).

I have a production system that has a large dataset of about 2 million rows. I need to create an index on a property but don't want it available to queries until the index is fully populated. Is there a way I can create the indexed, fire off the build, then "activate" the index so queries can use it.

The problem I face is the minute I create the index queries start returning zero rows because it starts using the index. As I start building it starts returning qualifying rows for those it has indexed but misses those not yet processed.

Hello Community,

As per the Build index documentation "If you use BUILD INDEX on a live system, the index is temporarily labeled as not selectable, meaning that queries cannot use the index while it is being built. Note that this will impact the performance of queries that use the index." Is this  hiding/not selectable is only applicable for BUILD INDEX or it supports class level %BuildIndices as well. as far as my analysis both syntax setting this setting  SetMapSelectability. 

Thanks!

Hi,

I know of the existance of (ELEMENTS) to create an index from a list, but I actually would like to index the content of an element of a list. Is it possible?

My scenario:

Class:
Property Test As list of TestList;
 

Test.List:
Property Name As %String;
Property Surname As %String;

I would like to have an  index based on the TestList.Name. If I try using

Index NewIndex On Test(ELEMENTS)

it will create an index with Name and Surname in it, but I just want to have an index with the name. Is it possible? 

Dear community!

I have problem with index NULL value.  Unique index doesn't work for this case.  If I use insert and one of parameter is "NULL". Message of constraint doesn't appear and row is inserted into table successfully.  How Can I use index with NULL?

Class TestClassIndx Extends %Persistent
{
Property name As %String [ Private ];
Property age As %String [ Private ];
Property country As %String [ Private ];
Index IndextestUniq On (name, age, country) [ Unique ];
}

INSERT INTO TestClassIndx (name,age,country) VALUES ('1','2',NULL)

Best Regards 

Is there any way to specify in which order indices are built during:

• %BuildIndices
• %Save

So far I think they are built by name order.

Hi-

I have the following objects

Class A

   Property P1 As B

   Property P2 As %String

   Property P3 As %String

Class B

   Property P1 As %String

Can I create an index in Class A based on P1.P1.  Basically I want an index of class A by property P1 in class B

I tried creating the following but got a compile error

Index I1 On P1.P1

Thanks

Hello Fellow Cache Developers:  

Has anyone ever created an index on values of a list property?   If so, would you be willing to share an example?

Also, feel free to offer input and suggestions regarding use of indexes on List values.

I have a class which defines a property as array of %String. Is it possible to index values of this property and use this property in SQL?

I have tried 'Index idx On prop(ELEMENTS)' and then a select from the generated collection table, but this is still orders of magnitude slower than queries to the containing class.

I have a persistent class that represents cities across the United States.  It is below, but basically has a City Id, Name, Lat, Lon and a few other unimportant fields for this issue.  Anytime I attempt to query on the Latitude or Longitude it immediately returns no results.  My first thought was that it was a casting issue so I tried casting both sides to floats, ints, even strings and in all cases it immediately comes back with no results.  I then decided to cast it to a string and attempt a like statement thinking it might be something about how floats are handled, but still no joy.

I have a unique index on two properties in a class, and I want to open the instance by index key with one of the two properties is set to the empty string. Is this possible to perform?

We have been storing raw messages in a MySQL database for DR and ad hoc purposes. We are thinking of using an Ensemble instance as our data lake instead. We could segregate the source data by namespace or by global. But either way we'll want a custom global to index the data for data retrieval performance purposes.

Anyone else taking this approach? Any feedback?