I'm running Node 8.9.4, Hapi 17.4, and Oracledb 2.2.
When attempting to call a stored procedure, I get the error "NJS-012: encountered invalid bind data type in parameter 2". Nothing I've been able to do seems to fix the issue. The code that calls the procedure is:
async function getSavedViews(req, h, server) {
let connection = await server.app.db.getConnection();
let bindVars = {
P_USER_NAME: req.payload.user_name,
P_CONTENT_TYPE: req.payload.content_type,
P_PROJECT_NUMBER: req.payload.project_number,
OP_GRID_TAB_TYP: { dir: server.app.db.BIND_OUT, type: server.app.db.ARRAY }
}
let res = server.methods.response();
try {
res.error = false;
res.msg = "Retrieved saved views.";
res.data = await connection.execute(
`BEGIN APPS.XXETA_GRID_USER_CONTEXT_PKG.EXTRACT_GRID_DETAILS(:P_USER_NAME, :P_CONTENT_TYPE, :P_PROJECT_NUMBER, :OP_GRID_TAB_TYP); END;`,
bindVars
);
} catch (err) {
server.app.logger.error(err.message);
res.error = true;
res.msg = err.message,
res.data = [];
}
return res;
}
The stored procedure is described as:
The error I get from my logger is: 2018-08-06 15:02:20 ERROR NJS-012: encountered invalid bind data type in parameter 2
Any help would be appreciated.
UPDATE:
The complex type that is the bound out variable looks like this...
CREATE OR REPLACE TYPE XXETA_GRID_CONTEXT_REC_TYP AS OBJECT
(
GRID_VIEW_ID NUMBER (15),
GRID_VIEW_NAME VARCHAR2 (240),
USER_NAME VARCHAR2 (30),
PROJECT_NUMBER VARCHAR2 (5)
)
Update 2019/08/28:
Node-oracledb added support for SQL object types and PL/SQL record types in v4 (released 2019/07/25). See this section of the doc for details: https://oracle.github.io/node-oracledb/doc/api.html#objects
Given the exact same objects as listed before, the following JavaScript can now be used to do the job with far fewer lines of code than before:
const oracledb = require('oracledb');
const config = require('./db-config.js');
async function runTest() {
let conn;
try {
const sql =
`call xxeta_grid_user_context_pkg.extract_grid_details(
p_user_name => :P_USER_NAME,
p_content_type => :P_CONTENT_TYPE,
p_project_number => :P_PROJECT_NUMBER,
op_grid_tab_typ => :OP_GRID_TAB_TYP
)`;
const binds = {
P_USER_NAME: 'Jane Doe',
P_CONTENT_TYPE: 'Some Content Type',
P_PROJECT_NUMBER: '123',
OP_GRID_TAB_TYP: {
dir: oracledb.BIND_OUT,
type: 'HR.XXETA_GRID_CONTEXT_TAB_TYP'
}
}
conn = await oracledb.getConnection(config);
const result = await conn.execute(
sql,
binds
);
const gridContexts = [];
for (let x = 0; x < result.outBinds.OP_GRID_TAB_TYP.length; x += 1) {
gridContexts.push({
gridViewId: result.outBinds.OP_GRID_TAB_TYP[x].GRID_VIEW_ID,
gridViewName: result.outBinds.OP_GRID_TAB_TYP[x].GRID_VIEW_NAME,
userName: result.outBinds.OP_GRID_TAB_TYP[x].USER_NAME,
projectNumber: result.outBinds.OP_GRID_TAB_TYP[x].PROJECT_NUMBER
});
}
console.log(gridContexts);
} catch (err) {
console.error(err);
} finally {
if (conn) {
try {
await conn.close();
} catch (err) {
console.error(err);
}
}
}
}
runTest();
Previous answer:
Complex types are not currently supported. The out bind you have specified falls in this category. Until such types are directly supported, you'll need to add a bit of wrapper code to break the complex type into one or more simple types. I show an example of this here: https://jsao.io/2017/01/plsql-record-types-and-the-node-js-driver/
The goal in that post is to invoke a stored procedure which accepts an array of a custom record type. To invoke it, I first have to declare some simple array types to bind into. Then I can use those arrays to create the more complex array and invoke the procedure.
In your case, you'll need to do the reverse. In the PL/SQL block, declare a local variable of type APPS.XXETA_GRID_CONTEXT_TAB_TYP. Then, after the procedure is invoked, iterate over the array and use it to populate some simple arrays (VARCHAR2, NUMBER, or DATE) and used those as your out binds.
Update:
Provided you have the following objects:
create or replace type xxeta_grid_context_rec_typ as object (
grid_view_id number(15),
grid_view_name varchar2(240),
user_name varchar2(30),
project_number varchar2(5)
)
/
create or replace type xxeta_grid_context_tab_typ as table of xxeta_grid_context_rec_typ
/
create or replace package xxeta_grid_user_context_pkg
as
procedure extract_grid_details(
p_user_name in varchar2,
p_content_type in varchar2,
p_project_number in varchar2,
op_grid_tab_typ out xxeta_grid_context_tab_typ
);
end;
/
create or replace package body xxeta_grid_user_context_pkg
as
procedure extract_grid_details(
p_user_name in varchar2,
p_content_type in varchar2,
p_project_number in varchar2,
op_grid_tab_typ out xxeta_grid_context_tab_typ
)
is
l_xxeta_grid_context_rec xxeta_grid_context_rec_typ;
begin
op_grid_tab_typ := xxeta_grid_context_tab_typ();
for x in 1 .. 3
loop
l_xxeta_grid_context_rec := xxeta_grid_context_rec_typ(
grid_view_id => x,
grid_view_name => 'Some Grid View',
user_name => p_user_name,
project_number => p_project_number
);
op_grid_tab_typ.extend();
op_grid_tab_typ(x) := l_xxeta_grid_context_rec;
end loop;
end;
end;
/
The following Node.js code can invoke the stored procedure and get the values from the complex out parameter.
const oracledb = require('oracledb');
const config = require('./dbConfig.js');
async function runTest() {
let conn;
try {
const userName = 'Jane Doe';
const contentType = 'Some Content Type';
const projectNumber = '123';
// This is what we want to populate with records/objects that come out
// of the procedure.
const gridContexts = [];
// We start by declaring some other arrays, one for each field in the
// xxeta_grid_context_rec_typ type.
const gridViewIds = [];
const gridViewNames = [];
const userNames = [];
const projectNumbers = [];
conn = await oracledb.getConnection(config);
// Then we execute the procedure with a little wrapper code to populate
// the individual arrays.
let result = await conn.execute(
`declare
-- This is a local variable that you'll use to get the out data from
-- the procedure.
l_xxeta_grid_context_tab xxeta_grid_context_tab_typ;
begin
xxeta_grid_user_context_pkg.extract_grid_details(
p_user_name => :user_name,
p_content_type => :content_type,
p_project_number => :project_number,
op_grid_tab_typ => l_xxeta_grid_context_tab
);
-- Now that the local variable is populated, iterate over it to
-- populate the individual out binds.
for x in 1 .. l_xxeta_grid_context_tab.count
loop
:grid_view_ids(x) := l_xxeta_grid_context_tab(x).grid_view_id;
:grid_view_names(x) := l_xxeta_grid_context_tab(x).grid_view_name;
:user_names(x) := l_xxeta_grid_context_tab(x).user_name;
:project_numbers(x) := l_xxeta_grid_context_tab(x).project_number;
end loop;
end;`,
{
user_name: userName,
content_type: contentType,
project_number: projectNumber,
grid_view_ids: {
dir: oracledb.BIND_OUT,
type: oracledb.NUMBER,
maxArraySize: 200
},
grid_view_names: {
dir: oracledb.BIND_OUT,
type: oracledb.STRING,
maxArraySize: 200
},
user_names: {
dir: oracledb.BIND_OUT,
type: oracledb.STRING,
maxArraySize: 200
},
project_numbers: {
dir: oracledb.BIND_OUT,
type: oracledb.STRING,
maxArraySize: 200
}
}
);
// At this point you can access the individual arrays to populate the
// original target array with objects. This is optional, you can work
// with the individual arrays directly as well.
for (let x = 0; x < result.outBinds.grid_view_ids.length; x += 1) {
gridContexts.push({
gridViewId: result.outBinds.grid_view_ids[x],
gridViewName: result.outBinds.grid_view_names[x],
userName: result.outBinds.user_names[x],
projectNumber: result.outBinds.project_numbers[x]
});
}
console.log(gridContexts);
} catch (err) {
console.error(err);
} finally {
if (conn) {
try {
await conn.close();
} catch (err) {
console.error(err);
}
}
}
}
runTest();
I hope that helps! Direct support for complex types is on the list of enhancements, just can't say when it will land.