I am planning to use NoSQL Cloud Service as our datastore. I have question about the MAP data type. Say I have a column “labels” ( labels MAP(RECORD(value STRING, contentType STRING)) in table “myTable”, which the “labels” column is MAP datatype and the value is RECORD data type .
I want to query the table which return all the rows that the key of the “labels” = particular value, what is the sql statement looks like? I tried:
select * from myTable where labels.keys($key=‘xxxx’)
which doesn’t work.
do we need to add the index for the label field in the MAP? any performance improvement? If yes, how to add this index?
Thanks
Please try the following syntax
select * from myTable t
where t.labels.keys() =any "xxx"
Your syntax is good if you add exists
select * from myTable t
where exists t.labels.keys($key= “xxx”)
Concerning your question about performance
there will be significant performance improvement.
If you want to index only the field names (keys) of the map,
you create the index like this:
create index idx_keys on myTable(labels.keys())
If you want to index both they keys and the associated values:
create index idx_keys_values
on myTable(labels.keys(), labels.values())
Related
I'm trying to figure out on how to create an Index for below query such that the SELECT statement only traverse the leaf level of the index horizontally and it does not access the relational table. I'm working on a relational database in Oracle.
SELECT SUM(SUM(qty))
FROM PlaceOrder
GROUP BY OrderNumber
HAVING COUNT(LineNumber) > 10;
Am I correct to create the below index?
CREATE INDEX IDX_PO
ON PlaceOrder(qty, OrderNumber, LineNumber);
Thank you.
As commented by astentx, not null constraint is needed. In fact, as long as any of qty, OrderNumber, LineNumber has not null, Oracle should be able to use the index.
Also, note that unless you specially want to exclude lines with null LineNumber, you can replace COUNT(LineNumber) with COUNT(OrderNumber) or even COUNT(*).
I have an oracle table. Table's DDL is (not have the primary key)
create table CLIENT_ACCOUNT
(
CLIENT_ID VARCHAR2(18) default ' ' not null,
ACCOUNT_ID VARCHAR2(18) default ' ' not null,
......
)
create unique index UK_ACCOUNT
on CLIENT_ACCOUNT (CLIENT_ID, ACCOUNT_ID)
Then, the data's scale is very huge, maybe 100M records. I want to traverse this whole table's data with batch.
Now, I use the table's index to batch traverse. But I have some oracle grammar problems.
# I want to use this SQL, but grammar error.
# try to use b-tree's index to locate start position, but not work
select * from CLIENT_ACCOUNT
WHERE (CLIENT_ID, ACCOUNT_ID) > (1,2)
AND ROWNUM < 1000
ORDER BY CLIENT_ID, ACCOUNT_ID
Has the fastest way to batch touch table data?
Wild guess:
select * from CLIENT_ACCOUNT
WHERE CLIENT_ID > '1'
and ACCOUNT_ID > '2'
AND ROWNUM < 1000;
It would at least compile, although whether it correctly implements your business logic is a different matter. Note that I have cast your filter criteria to strings. This is because your columns have a string datatype and you are defaulting them to spaces, so there's a high probability those columns contain non-numeric values.
If this doesn't solve your problem, please edit your question with more details; sample input data and expected output is always helpful in these situations.
Your data model seems odd.
Your columns are defined as varchar2. So why is your criteria numeric?
Also, why do you default the key columns to space? It would be better to leave unpopulated values as null. (To be clear, NULL is not a good thing in an indexed column, it's just better than a space.)
I have a question about the MAP data type. Say I have a column labels ( labels MAP(RECORD(value STRING, contentType STRING)) in myTable, which the “labels” column is MAP data type and the value is a RECORD data type .
I want to query the table which returns all the rows that the key of the "labels" "startsWith" particular value ("xxx.*"),
I've tried this but I am wondering if there is a better way to do
Select labels.keys($key >='xxx') as keys,
labels.values($key >='xxx') as values
from myTable where labels.keys() >=any ('xxx')
You can try
select * from myTableName t
where exists t.labels.keys(starts_with($key, 'xxx'));
or
select f.labels.keys(regex_like($key,'xxx.*')) as keys,
f.labels.values(regex_like($key,'xxx.*')) as values
from myTable f
I also suggest changing from MAP to ARRAY, which can support path filter to get the matched entries. In the previous examples, the order between the values and keys is not guaranteed
select labels[regex_like($element.label ,‘xxx.*’)] from myTable
I want to write a function that returns table records that I can show in my app's grid. My app shall be oblivious to the table's structure. I want it to be still working when someone adds columns to the table or removes them.
In my app it shall look something like this:
myQuery = "select * from table(myfunction)";
...
The function should hence be something like:
CREATE OR REPLACE FUNCTION myfunction RETURN TABLE OF mytable%ROWTYPE AS ...
But RETURN TABLE OF mytable%ROWTYPE is not allowed. One has to create an SQL type.
But CREATE TYPE table_of_mytable_rows IS TABLE OF mytable%ROWTYPE is not allowed either.
So, is there a way to achieve exactly what I want?
As an alternative I thought of working with IDs. This would make my app's code look something like this:
myQuery = "select * from mytable where rowid in (select * from table(myfunction))";
...
And the function would then be
CREATE OR REPLACE FUNCTION myfunction RETURN TABLE OF UROWID AS ...
But again, RETURN TABLE OF UROWID is not allowed. Neither is creating the SQL table type with CREATE TYPE table_of_rowids IS TABLE OF UROWID.
I know I could create an object type resembling the table's primary key columns and then create a table type on this object. Then my app would have to know the table's primary key in order to
myQuery = "select * from mytable where (key1, key2) in (select key1, key2 from table(myfunction))";
...
I would much prefer my function to return table rows or rowids. Is this possible?
I'm currently still working on Oracle 11.2, but I would also be interested in solutions for newer versions.
I need to search rows entered on a specific date.
However the datatype of column I need to search on is datetime, and the datatype of argument is Date.
I can use the the query like
Select result
from table
where
convert(date, Mycolumn) = #selectedDate
but this would affect the SARGability of the query and will not use indexes created on mycolumn.
I was trying to use the following query:
Select result
from table
where
Mycolumn
BETWEEN #selectedDate AND Dateadd(s, -1, Dateadd(D, 1, #selectedDate))
However this does not work since the #selectedDate is Date type and a second can't be added or removed.
Can someone help me with a working query?
Thanks.
It is my understanding that using:
convert(date, Mycolumn) = #selectedDate
is SARGable. It will use the index on Mycolumn (if one exists). This can easily be confirmed by using the execution plan.
Select result
from table
where
Mycolumn >= #selectedDate
AND Mycolumn < Dateadd(D, 1, #selectedDate)
If you need to do these searches a lot, you could add a computed, persisted column that does the conversion to DATE, put an index on it and then search on that column
ALTER TABLE dbo.YourTable
ADD DateOnly AS CAST(MyColumn AS DATE) PERSISTED
Since it's persisted, it's (re-)calculated only when the MyColumn value changes, e.g. it's not a "hidden" call to a stored function. Since it's persisted, it can also be indexed and used just like any other regular column:
CREATE NONCLUSTERED INDEX IX01_YourTable_DateOnly ON dbo.YourTable(DateOnly)
and then do:
SELECT result FROM dbo.YourTable WHERE DateOnly = #SelectedDate
Since that additional info is stored in the table, you'll be using a bit more storage - so you're doing the classic "space vs. speed" trade-off; you need a bit more space, but you get more speed out of it.