We Keep Coding

Can't delete a row previously created with an upsert in Cassandra using Java

the TL;DR is that I am not able to delete a row previously created with an upsert using Java.
Basically I have a table like this:
CREATE TABLE transactions (
key text PRIMARY KEY,
created_at timestamp
);
Then I execute:
String sql = "update transactions set created_at = toTimestamp(now()) where key = 'test' if created_at = null";
session.execute(sql)
As expected the row is created:
cqlsh:thingleme> SELECT * FROM transactions ;
key | created_at
------+---------------------------------
test | 2018-01-30 16:35:16.663000+0000
But (this is what is making me crazy) if I execute:
sql = "delete from transactions where key = 'test'";
ResultSet resultSet = session.execute(sql);
Nothing happens. I mean: no exception is thrown and the row is still there!
Some other weird stuff:
if I replace the upsert with a plain insert, then the delete works
if I directly run the sql code (update and delete) by using cqlsh, it works
If I run this code against an EmbeddedCassandraService, it works (this is very bad, because my integration tests are just green!)
My environment:
cassandra: 3.11.1
datastax java driver: 3.4.0
docker image: cassandra:3.11.1
Any idea/suggestion on how to tackle this problem is really appreciated ;-)

I think the issue you are encountering might be explained by the mixing of lightweight transactions (LWTs) (update transactions set created_at = toTimestamp(now()) where key = 'test' if created_at = null) and non-LWTs (delete from transactions where key = 'test').
Cassandra uses timestamps to determine which mutations (deletes, updates) are the most recently applied. When using LWTs, the timestamp assignment is different then when not using LWTs:
Lightweight transactions will block other lightweight transactions from occurring, but will not stop normal read and write operations from occurring. Lightweight transactions use a timestamping mechanism different than for normal operations and mixing LWTs and normal operations can result in errors. If lightweight transactions are used to write to a row within a partition, only lightweight transactions for both read and write operations should be used.
Source: How do I accomplish lightweight transactions with linearizable consistency?
Further complicating things is that by default the java driver uses client timestamps, meaning the write timestamp is determined by the client rather than the coordinating cassandra node. However, when you use LWTs, the client timestamp is bypassed. In your case, unless you disable client timestamps, your non-LWT queries are using client timestamps, where your LWT queries are using a timestamp assigned by the paxos logic in cassandra. In any case, even if the driver wasn't assigning client timestamps this still might be a problem because the timestamp assignment logic is different on the C* side for LWT and non-LWT as well.
To fix this, you could alter your delete statement to include IF EXISTS, i.e.:
delete from transactions where key = 'test' if exists
Similar issue from the java driver mailing list

Java PreparedStatement Cross-DB with casting

I have a PreparedStatement intended to be run both on ORACLE and on MYSQL.
But I cannot figure out how to handle the CAST(NULL AS ...)
On Oracle the following works (but not on Mysql):
SELECT TIMB_INS,
CAST(NULL AS TIMESTAMP) AS TIMB_CLO
FROM TOPS
On Mysql the following works (but not on Oracle):
SELECT TIMB_INS,
CAST(NULL AS DATETIME) AS TIMB_CLO
FROM TOPS
(Please note that the first column selected, "TIMB_INS", returns the correct data type for target database type in both cases, i.e. TIMESTAMP for Oracle and DATETIME for MySql.)
There is a way to put it so that it works for both?
I.E. Can i make it db-indipendent in some way?
Thanks
Marco

Based on the tags I can see you're calling this statement from some java code. There are several ways doing so:
Use the DAO pattern. I.e. for each SQL flavor provide a java file that contains the SQL-s.
Use an ORM like Hibernate or JPA. That will take care of this kind of differences.
As a quick hack, you can edit the SQL manually, like in the snippet below. But then you have to determine somehow if the underlying database is Oracle or MySQL
String SQL_PATTERN = "... CAST(NULL AS %s) AS TIMB_CLO ...";
String SQL = String.format(SQL_PATTERN, isOracle ? "TIMESTAMP" : "DATETIME");

Change Table names in derby database using entitymanager

I am using an APACHE DERBY database, and basing my database interactions on EntityManager, and I don't want to use JDBC class to build a query to change my tables' names (i just need to put a prefix to each new user to the application, but have the same structure of tables), such as:
//em stands for EntityManager object
Query tableNamesQuery= em.createNamedQuery("RENAME TABLE SCHEMA.EMP_ACT TO EMPLOYEE_ACT");
em.executeUpdate();
// ... rest of the function's work
// The command works from the database command prompt but i don't know how to use it in a program
//Or as i know you can't change system tables data, but here's the code
Query tableNamesQuery= em.createNamedQuery("UPDATE SYS.SYSTABLES SET TABLENAME='NEW_TABLE_NAME' WHERE TABLETYPE='T'");
em.executeUpdate();
// ... rest of the function's work
My questions are :
This syntax is correct?
Will it work?
Is there any other alternative?
Should I just use the SYS.SYSTABLES and find all the tables that has 'T' as tabletype and alter their name their, will it change the access name ?

I think you're looking for the RENAME TABLE statement: http://db.apache.org/derby/docs/10.10/ref/rrefsqljrenametablestatement.html
Don't just issue update statements against the system catalogs, you will corrupt your database.

What is the correct usage of zxjdbc to call stored procedures?

I am attempting to use zxJDBC to connect to a database running on SQL Server 2008 R2 (Express) and call a stored procedure, passing it a single parameter. I am using jython-standalone 2.5.3 and ideally do not want to have to install additional modules.
My test code is shown below.
The database name is CSM
Stored Procedure:
SET ANSI_NULLS ON
GO
SET QUOTED_IDENTIFIER ON
GO
-- =============================================
-- Author: <Author,,Name>
-- Create date: <Create Date,,>
-- Description: <Description,,>
-- =============================================
CREATE PROCEDURE dbo.DUMMY
-- Add the parameters for the stored procedure here
#carrierId VARCHAR(50)
AS
BEGIN
-- SET NOCOUNT ON added to prevent extra result sets from
-- interfering with SELECT statements.
SET NOCOUNT ON;
-- Insert statements for procedure here
INSERT INTO dbo.carrier (carrierId, test)
VALUES (#carrierId, 'Success')
END
GO
Jython Script:
from com.ziclix.python.sql import zxJDBC
conn = None
try :
conn = zxJDBC.connect('jdbc:sqlserver://localhost\SQLEXPRESS', 'sa', 'password', 'com.microsoft.sqlserver.jdbc.SQLServerDriver')
cur = conn.cursor()
cur.callproc(('CSM','dbo','DUMMY'), ['carrier1'])
conn.commit()
except Exception, err :
print err
if conn:
conn.rollback()
finally :
if conn :
conn.close()
By using cur.execute() I have been able to verify that the above is successfully connecting to the database, and that I can query against it. However, I have thus far been unable to successfully call a stored procedure with parameters.
The documentation here(possibly out of date?) indicates that callproc() can be called with either a string or a tuple to identify the procedure. The example given -
c.callproc(("northwind", "dbo", "SalesByCategory"), ["Seafood", "1998"], maxrows=2)
When I attempt to use this method, I receive the following error
Error("Could not find stored procedure 'CSM.DUMMY'. [SQLCode: 2812], [SQLState: S00062]",)
It would appear that zxJDBC is neglecting to include the dbo part of the procedure identifier.
If I instead call callproc with "CSM.dbo.DUMMY" as the first argument then I receive this error
Error('An object or column name is missing or empty. For SELECT INTO statements, verify each column has a name. For other statements, look for empty alias names. Aliases defined as "" or [] are not allowed. Change the alias to a valid name. [SQLCode: 1038], [SQLState: S0004]',)
Using a profiler on the database whilst running my script shows that in the second case the following SQL is executed:
use []
go
So it would seem that when using a single string to identify the procedure, the database name is not correctly parsed out.
One of my trial and error attempts to fix this was to call callproc as follows:
cur.callproc(('CSM', '', 'dbo.DUMMY'), ['carrier1'])
This got me only as far as
Error("Procedure or function 'DUMMY' expects parameter '#carrierId', which was not supplied. [SQLCode: 201], [SQLState: S0004]",)
In this case what I think is happening is that zxJDBC attempts to call a system stored procedure (sp_proc_columns) to determine the required parameters for the stored procedure I want to call. My guess is that with the procedure identifier in the incorrect format above, zxJDBC does not get a valid/correct return and assumes no parameters are required.
So basically I am not a bit stuck for ideas as to how to get it to
Use the correct database name
Correctly determine the required parameters using sp_proc_columns
Call my stored procedure with the correct name
all at the same time.
I do have a workaround, which is to use something like
cur.execute('EXEC CSM.dbo.DUMMY ?', ['carrier1'])
However I feel like callproc() is the correct solution, and would likely produce cleaner code when I come to call stored procedures with large numbers of parameters.
If anyone can spot the mistake(s) that I am making, or knows that this is not ever going to work as I think then any input would be much appreciated.
Thanks
Edit
As suggested by i-one, I tried adding cur.execute('USE CSM') before calling my stored procedure (also removing the database name from the procedure call). This unfortunately produces the same Object or Column missing error as above. The profiler shows USE CSM being executed, followed by USE [] so it seems that callproc() always fires a USE statement before the procedure itself.
I have also experimented with turning on/off autocommit, to no avail.
Edit 2
Further information following comments/suggested solutions:
"SQLEXPRESS" in my connection string is the database instance name.
Using double quotes instead of single has no effect.
Including the database name in the connection string (via ;databaseName=CSM; as specified here) and omitting it from the callproc() call leads to the original error with a USE [] statement being fired.
Using callproc(('CSM', 'dbo', 'dbo.DUMMY'), ['carrier1']) gives me some progress but results in the error
Error("Procedure or function 'DUMMY' expects parameter '#carrierId', which was not supplied. [SQLCode: 201], [SQLState: S0004]",)
I'll attempt to investigate this further
Edit 3
Based on the queries I could see zxJDBC firing, I manually executed the following against my database:
use CSM
go
exec sp_sproc_columns_100 N'dbo.DUMMY',N'dbo',N'CSM',NULL,N'3'
go
This gave me an empty results set, which would seem to explain why zxJDBC isn't passing any parameters to the stored procedure - it doesn't think it needs to. I have yet to figure out why this is happening though.
Edit 4
To update the above, the empty result set is because the call should be
exec sp_sproc_columns_100 N'DUMMY',N'dbo',N'CSM',NULL,N'3'
This unfortunately brings me full circle as I can't remove the dbo owner from the stored procedure name in my callproc() call or the procedure won't be found at all.
Edit 5
Table definition as requested
CREATE TABLE [dbo].[carrier](
[carrierId] [varchar](50) NOT NULL,
[test] [varchar](50) NULL
) ON [PRIMARY]

Though completely unaware of the technologies used here (unless some minor knowledge of SQL Server), I will attempt an answer (please forgive me if my jython syntax is not correct. I am trying to outline possibilities here not exact code)
My first approach (found at this post) would be to try:
cur.execute("use CSM")
cur.callproc(("CSM","dbo","dbo.DUMMY"), ["carrier1"])
This must have to do with the fact that sa users always have the dbo as a default schema (described at this SO post)
If the above does not work I would also try to use the CSM database name in the JDBC url (this is very common when using JDBC for other databases) and then simply call one of the two below.
cur.callproc("DUMMY", ["carrier1"])
cur.callproc("dbo.DUMMY", ["carrier1"])
I hope this helps
Update: I quote the relevant part of the link that you can't view
>> Program calls a Stored Procedure - master.dbo.xp_fixeddrives on MS SQL Server
from com.ziclix.python.sql import zxJDBC
def getConnection():
url = "${DBServer.Url}"
user= "${DBServer.User}"
password = "${DBServer.Password}"
driver = "${DBServer.Driver}"
con = zxJDBC.connect(url, user, password, driver)
return con
try:
conn = getConnection()
print 'Connection successful'
cur = conn.cursor()
cur.execute("use master")
cur.callproc(("master", "dbo", "dbo.xp_fixeddrives"))
print cur.description
for a in cur.fetchall():
print a
finally:
cur.close()
conn.close()
print 'Connection closed'
The error you get when you specified the call function like above suggests that the parameter is not passed correctly. So please modify your stored procedure to take a default value and try to call with passing params = [None]. If you see that the call succeeds we must have done something right as far as specifying the database is concerned.
Btw: the most recent documentation suggests that you should be able to access it with your syntax.

As outlined in comments callproc will work only with SELECT. Try this approach instead:
cur.execute("exec CSM.dbo.DUMMY #Param1='" + str(Param1) + "', #carrierId=" + str(carrierID))
Please see this link for more detail.

Apache Camel example to insert a row in a table

I want to insert exchange.body to a database table for one of the condition of my route.
Is there any example/tutorial of camel-jdbc component to insert message body?
Can I import the SQL statement itself and pass exchange.body to it?
I looked at http://camel.apache.org/jdbc.html example, but could not understand it.
Here Spring example is confusing for me. I didn't get why is it setting the body as SQL query and again importing some query from the class path. (There is no insert query example mentioned here.)

If you want to insert using the same statement (changing the parameters only) - use SQL component.
If you want to insert using arbitrary SQL statement into the component - use JDBC component.
SQL component usage:
from("direct:start").to("sql:insert into table foo (c1, c1) values ('#','#')");
com.google.common.collect.Lists;
producerTemplate.sendBody("direct:start", Lists.newArrayList("value1","value2"));
JDBC component usage:
from("direct:start").to("jdbc:dataSource");
producerTemplate.sendBody("direct:start", "insert into table foo (c1, c1) values ('value1','value2')");

You probably need to do some restructure of your payload before inserting it anyway, so there should probably be no issue to do a transformation using whatever method in Camel to set the body to the appropriate INSERT statement.
The important thing is what kind of payload structure your incoming message have. In the basic case - it's a string - it should be fairly simple
// In a Java bean/processor before the JDBC endpoint.
// Update: make sure to sanitize the payload from SQL injections if it contains user inputs or external data not generated by trusted sources.
exchange.getIn().setBody("INSERT INTO MYTABLE VALUES('" + exchange.getIn().getBody(String.class) + "', 'fixedValue', 1.0, 42)");
In case your message contains complex data structures, this code will of course be more complex, but it's pretty much the same way regular application will generate SQL queries.
The classpath example you are refering to
<jdbc:embedded-database id="testdb" type="DERBY">
<jdbc:script location="classpath:sql/init.sql"/>
</jdbc:embedded-database>
Simply shows how to test the JDBC component by starting a Database server embedded (Apache Derby) and populate it with some initial data (the sql/init.sql file). This part is not really part of the core jdbc component, but simply in the documentation to get up and running a sample without needing to configure a DB server and setup the JDBC connection properties.
That said, you might want to use the SQL component for more complex scenarios.