Develop Reference

JPA date truncation group by using POSTGRES and eclipselink

I am trying to truncate a date, and group by the values.
I have tried this:
JPA
select v.sop, FUNC('DATE_TRUNC', 'day', v.scrappedAt) as dt, sum(v.totalValue)
from TABLE v
where v.coordStatus like 'done%' and (:plant is null or v.target =
:plant) and v.scrappedAt is not null
group by v.sop, dt
I have also tried grouping by the FUNC
select v.sop, FUNC('DATE_TRUNC', 'day', v.scrappedAt) as dt, sum(v.totalValue)
from TABLE v
where v.coordStatus like 'done%' and (:plant is null or v.target =
:plant) and v.scrappedAt is not null
group by v.sop, FUNC('DATE_TRUNC', 'day', v.scrappedAt)
The error is the same
org.postgresql.util.PSQLException:
ERROR: column "t_mcp_verschrottungs_db_sharepoint.scrappedat"
must appear in the GROUP BY clause or be used in an aggregate function

Strictly speaking, if FUNC were a native Postgres function, then your query should be working, and the query would be ANSI compliant. After reading this SO question, it appears that Postgres can't figure out that the two FUNC calls are actually the same thing. Perhaps you can rephrase the sum using a correlated subquery:
SELECT
v1.sop,
FUNC('DATE_TRUNC', 'day', v1.scrappedAt) AS dt,
(SELECT SUM(totalValue) FROM TABLE v2
WHERE v2.sop = v1.sop AND
FUNC('DATE_TRUNC', 'day', v2.scrappedAt) =
FUNC('DATE_TRUNC', 'day', v1.scrappedAt) AND
v2.acoordStatus LIKE 'done%' AND
(:plant is null OR v2.target = :plant) AND
v2.scrappedAt IS NOT NULL) value_sum
FROM TABLE v1
WHERE
v1.coordStatus LIKE 'done%' AND
(:plant is null OR v1.target = :plant) AND
v1.scrappedAt IS NOT NULL;
An alternative to the above, should it either not work or not be performant, would be to use a native Postgres query.

Get primary keys of updated rows when doing an update with jdbi

I'm using jdbi (but would prepared to use raw jdbc if needed). My DB is currently Oracle. I have an update that updates the first row matching certain criteria. I want to get the primary key of the updated row from the same statement. Is this possible?
I tried
Integer rowNum = handle
.createUpdate(sqlFindUpdate)
.bind("some var", myVal)
.executeAndReturnGeneratedKeys("id")
.mapTo(Integer.class)
.findOnly();
but I guess this isn't a generated key, as it doesn't find it (illegal state exception, but the update succeeds).
Basically, I have a list of items in the DB that I need to process. So, I want to get the next and mark it as "in progress" at the same time. I'd like to be able to support multiple worker threads, so it needs to be a single statement - I can't do the select after (the status has changed so it won't match anymore) and doing it before introduces a race condition.
I guess I could do a stored procedure that uses returning into but can I do it directly from java?

I'm answering my own question, but I don't think it's a good answer :) What I'm doing is kind of a hybrid. It is possible to dynamically run PL/SQL blocks from jdbi. Technically, this is from Java as I had asked, not via a stored procedure. However, it's kind of a hack, in my opinion - in this case why not just create the stored procedure (as I probably will, if I don't find a better solution). But, for info, instead of:
String sql = "update foo set status = 1 where rownr in (select rownr from (select rownr from foo where runid = :runid and status = 0 order by rownr) where rownum = 1)";
return jdbi.withHandle((handle) -> {
handle
.createUpdate(sql)
.bind("runid", runId)
.executeAndReturnGeneratedKeys("rownr")
.mapTo(Integer.class)
.findOnly();
});
you can do
String sql = "declare\n" +
"vRownr foo.rownr%type;\n" +
"begin\n" +
"update foo set status = 1 where rownr in (select rownr from (select rownr from foo where runid = :runid and status = 0 order by rownr) where rownum = 1) returning rownr into vRownr;\n" +
":rownr := vRownr;\n" +
"end;";
return jdbi.withHandle((handle) -> {
OutParameters params = handle
.createCall(sql)
.bind("runid", runId)
.registerOutParameter("rownr", Types.INTEGER)
.invoke();
return params.getInt("rownr");
});
Like I said, it's probably better to just create the procedure in this case, but it does give you the option to still build the SQL dynamically in java if you need to I guess.
Based on this question, as linked by #APC in the comments, it is possible to use the OracleReturning class without the declare/begin/end.
String sql = "update foo set status = 1 where rownr in (select rownr from (select rownr from foo where runid = ? and status = 0 order by rownr) where rownum = 1) returning rownr into ?";
return jdbi.withHandle((handle) -> {
handle
.createUpdate(sql)
.bind(0, runId)
.addCustomizer(OracleReturning.returnParameters().register(1, OracleTypes.INTEGER))
.execute(OracleReturning.returningDml())
.mapTo(Integer.class)
.findOnly();
});
However, OracleReturning doesn't support named parameters, so you have to use positionals. Since my main reason for using JDBI over plain JDBC is to get named parameter support, that's important to me, so I'm not sure which way I'll go
Pretty hard dependency on it being an Oracle DB you're calling...
Update: enhancement for named parameters in OracleReturning was merged to master, and will be included in 3.1.0 release. Kudos to #qualidafial for the patch

JDBC: multicolumn IN query

I have a following query:
SELECT
date, userId, value
FROM
tbl_table
WHERE
date = to_date(:date, 'YYYY-MM-DD')
AND
userId = :userId
It allows to request for a single value like this:
MapSqlParameterSource args = new MapSqlParameterSource();
args.addValue("date", date, Types.VARCHAR);
args.addValue("userId", userId, Types.VARCHAR);
SqlRowSet rowSet = jdbcTemplate.queryForRowSet(SQL_SELECT, args);
jdbcTemplate.queryForRowSet(SQL_SELECT_MARKET_VALUE, args);
This is totally ok, but extremelly slow in case you have to query value for many date/userId pairs.
I would like to optimize it using multicolumn IN clause, but how do I handle multicolumn list via JDBC (or better question: is it possible using JDBC)?

Oracle supports multiple columns in "in" predicate:
SELECT
date, userId, value
FROM
tbl_table
WHERE
(date, userId) IN ((to_date(:date1, 'YYYY-MM-DD'), :userId1), (to_date(:date2, 'YYYY-MM-DD'), :userId2))
However JDBC doesn't provide a decent support of in-statement parameters - you will have to build the query using StringBuilder or use some of workarounds described here

It depends of details. If user/date filter is quite persistent (should be user more than once) temporary table will be the best decision. You can fill it once, you can edit it, and you can use it several times without reloading.
If you need of quite large number of pairs, I'd recommend you to use a table type. It would be something like this:
create type DateUserPair as object (dt date, userid integer);
create type DateUserPairs as table of DateUserPair;
....
SELECT
date, userId, value
FROM
tbl_table src,
table(cast :filter as DateUserPairs) flt
WHERE
src.date = flt.dt and
src.userId = flt.userId;
If filter would be small, filtering by (date, userId) in ((?,?), (?,?), ...) would be simple and clever.
Btw, your approach
date = to_date(:date, 'YYYY-MM-DD')
isn't good practise. Such conversions should be done by client, not by server. Use
date = :date
and assign it as date instead.

If what you want is to pass JDBC a list of date/userId pairs, or a list of dates and a list of userIds, I think it will not work.
A possible workaround in Oracle would be using a global temporary table with ON COMMIT DELETE ROWS. Your would have:
-- DDL for the workaround
CREATE GLOBAL TEMPORARY TABLE admin_work_area
(d DATE,
userId VARCHAR2(10))
ON COMMIT DELETE ROWS;
...
-- Start of query method pseudo-code
...
-- You should be able to JDBC-batch these for better performance
INSERT INTO temp_multicolumn_filter (d, userId) VALUES (date1, userId1);
INSERT INTO temp_multicolumn_filter (d, userId) VALUES (date2, userId2);
...
-- Query using temp_multicolumn_filter
SELECT date, userId, value
FROM tbl_table
WHERE
(date, userId) in (select d, userId from temp_multicolumn_filter);
...
-- End of query method pseudo-code
As the temporary table has the ON COMMIT DELETE ROWS, each transaction will only see its own date/userId pairs. Just remember that if you use the temporary table more than once in the same transaction, you might need to clear it before using it.
UPDATE:
Another option would be using a PIPELINED PL/SQL function to "build" your table inside the query:
-- DDL for this workaround
CREATE TYPE date_userid_pair AS OBJECT (
d DATE,
userId VARCHAR2(10));
CREATE TYPE date_userid_dataset IS TABLE OF date_userid_pair;
CREATE FUNCTION decode_date_userid_pairs(dates in varchar2, userIds in varchar2)
RETURN date_userid_dataset PIPELINED IS
result_row date_userid_pair;
BEGIN
WHILE there are more "rows" in the parameters LOOP
result_row.d := -- Decode next date from dates
result_row.userId := -- Decode next userId from userIds
PIPE ROW(result_row);
END LOOP;
END;
// Start of query method pseudo-code
...
// This is Java code: encodeList encodes a List of elements into a String.
encodedDates = encodeList(listOfDates);
encodedUserIds = encodeList(listOfUserIds);
...
// Query using temp_multicolumn_filter
SELECT date, userId, value
FROM tbl_table
WHERE
(date, userId) in (
select date, userId
from TABLE(decode_date_userid_pair(:encodedDates, :encodedUserIds));
...
// End of query method pseudo-code
But this is more hacky, and if you don't have privileges to create a temporary table, then you probably won't have CREATE TYPE either (you might not even have CREATE FUNCTION privilege).

extract column names from sql query after where clause

I've a requirement where I need to pull out data from database.
The query is-
SELECT e.Data AS EntityBlob, f.Data AS FpmlBlob
FROM [Trades.InventoryRecord] ir, EntityBlob e, FpmlBlob f
WHERE %s AND uid = e.uid AND uid = f.uid
Here %s is the predicate after where clause which user will input from an html form.
User input will be in this form :
1. TradeDate = '2013-04-05' AND IsLatest = 'TRUE'
2. StreamId= 'IA0015'
3. The query may have IN clause also
Now when this query is rendered I get exception ambigous column streamId or ambigous column IsLatest, as these columns exists in more than one table with same name. So to remove this ambiguity I need to modify the query as - ir.IsLatest or ir.StreamId
To do so by java code, I need to first parse the predicate after where clause, extract column names and insert table name alias- 'ir' before each column name so that the query becomes -
SELECT e.Data AS EntityBlob, f.Data AS FpmlBlob
FROM [Trades.InventoryRecord] ir, EntityBlob e, FpmlBlob f
WHERE ir.TradeDate = '2013-04-05' AND ir.IsLatest = 'TRUE' AND uid = e.uid AND uid = f.uid
what is the best way to parse this predicate, or if there is any other way I can achieve the same result?

My answer to this question is to not parse the user input - there is far too much that can go wrong. It would be a lot better to have a UI with drop downs and buttons for selecting equality, inequality, ranges, in statements, etc. It may seem like more work, but protecting yourself from a SQL injection attack is even more. And even if you are not concerned about malicious SQL injection, then the user still has to get every thing exactly right, or the statement fails.

How to construct advanced Hibernate Query with OR and summarize functions

I have a rather complex query which works in SQL, but I would like to express this in HQL for portability. I'm going to fetch a user configured preference value if they exist, if not I must use a default value. This value must be subtracted from current date and the matched against a column in the table which I'm interested in:
select d.id, d.guid, d.deletetimestamp, u.id
from descriptor d, ownerkey ow, user u
where
d.parentid in
(select td.id
from descriptor td, store s
where s.type = 'Trash'
and s.descriptorid = td.id
)
and d.ownerkeyid = ow.id
and ow.ownerid = u.id
and
(
(d.deletetimestamp < CURRENT_TIMESTAMP() - INTERVAL
(select pv.value
from preferencevalue pv, userpreference up
where u.id = up.userid
and up.preferenceid = 26
and up.value = pv.id)
DAY)
or
(d.deletetimestamp < CURRENT_TIMESTAMP() - INTERVAL
(select cast(pv.value as SIGNED)
from preferencevalue pv, defaultpreference dp
where dp.preferenceid = 26
and not exists(select up.userid from userpreference up where u.id = up.userid and up.preferenceid = 26)
and dp.value = pv.id)
DAY)
)
I'm trying to construct this by using the Criteria API which seems to include most of the logical operators that I need (equals, larger than, or, isEmpty/isNull), but not sure how I would express all these parts.
Using a view is not an option at this point since we're using MySQL as the production database while the integration tests are running with H2 inmemory database. I'm not able to get find the sata substract function in H2 while MySQL do support this.
The select fields isn't important since they have only been used for testing purposes.

you can use Restrictions.disjunction() for or -and Restrictions.conjuction() for and clauses.
To reference a certain property of an entity (like pv.value) you can use Projections.property("value")
for the casting I'm not sure, perhaps using the #Formula annotation on your entity? But this is a hibernate and not a JPA annotation.
as far as I know there is no equivalent for INTERVAL in hibernate but in such cases (maybe also for the above cast) you could use Restrictions.sqlRestriction("some sql...")
It will be a challenge putting all of this together to transform your query to hibernate criteria.
greetz,
Stijn