OBIEE BI Answers: Wrong Aggregation Measures on top level of hierarchy

Question

Hi to all,
I have following problem. I hope to be clear in my English because it's a bit complicated to explain.
I have following fact table:
Drug Id Ordered Quantity
1 9
2 4
1 3
2 2
and following Drug Table:
Drug Brand Id Brand Description Drug Active Ingredient Id Drug Active Ingredient Description
1 Aulin 1 Nimesulide
2 Asprina 2 Acetilsalicilico
In AWM i've defined a Drug Dimension based on following hierarchy: Drug Active Ingredient (parent) - Drug Brand Description (leaf) mapped as:
Drug Active Ingredient = Drug Active Ingredient Id of my Drug Table (LONG DESCRIPTION Attribute=Drug Active Ingredient Description)
Drug Brand Description = Drug Brand Id of my Drug Table (LONG DESCRIPTION Attribute = Drug Brand Description)
Indeed in my cube I've mapped leaf level Drug Brand Description = Drug Id of my fact table. In AWM Drug Dimension is mapped as Sum Aggregation Operator
If I select on Answers Drug Active Ingredient (parent of my hierarchy) and Ordered Quantity I see following result
Drug Active Ingredient Description Ordered Quantity
Acetilsalicilico 24
Nimesulide 12
indeed of correct values
Drug Active Ingredient Description Ordered Quantity
Acetilsalicilico 12
Nimesulide 6
EXACTLY the double!!!!!!! But if I drill down Drug Active Ingredient Description Acetilsalicilico I see correctly:
Drug Active Ingredient Description Drug Brand Description Ordered Quantity
Acetilsalicilico
- Aspirina 12
Total 12
Wrong Aggregation is only on top level of hierarchy. Aggregation on lower level of hierarchy is correct. Maybe Answers sum also Total Row????? Why?????
I'm frustrated. I beg your help, please!!!!!!!!
Giancarlo

Answer 1

Hi,
in NQSConfig.ini I can't find Cache Section. I post all file. Tell me what I must change. I know your patient is quite at limit!!!!!!! But I'm a new user of OBIEE.
# NQSConfig.INI
# Copyright (c) 1997-2006 Oracle Corporation, All rights reserved
# INI file parser rules are:
# If values are in literals, digits or _, they can be
# given as such. If values contain characters other than
# literals, digits or _, values must be given in quotes.
# Repository Section
# Repositories are defined as logical repository name - file name
# pairs. ODBC drivers use logical repository name defined in this
# section.
# All repositories must reside in OracleBI\server\Repository
# directory, where OracleBI is the directory in which the Oracle BI
# Server software is installed.
[ REPOSITORY ]
#Star     =     samplesales.rpd, DEFAULT;
Star = Step3.rpd, DEFAULT;
# Query Result Cache Section
[ CACHE ]
ENABLE     =     YES;
// A comma separated list of <directory maxSize> pair(s)
// e.g. DATA_STORAGE_PATHS = "d:\OracleBIData\nQSCache" 500 MB;
DATA_STORAGE_PATHS     =     "C:\OracleBIData\cache" 500 MB;
MAX_ROWS_PER_CACHE_ENTRY = 100000; // 0 is unlimited size
MAX_CACHE_ENTRY_SIZE = 1 MB;
MAX_CACHE_ENTRIES = 1000;
POPULATE_AGGREGATE_ROLLUP_HITS = NO;
USE_ADVANCED_HIT_DETECTION = NO;
MAX_SUBEXPR_SEARCH_DEPTH = 7;
// Cluster-aware cache
// GLOBAL_CACHE_STORAGE_PATH = "<directory name>" SIZE;
// MAX_GLOBAL_CACHE_ENTRIES = 1000;
// CACHE_POLL_SECONDS = 300;
// CLUSTER_AWARE_CACHE_LOGGING = NO;
# General Section
# Contains general server default parameters, including localization
# and internationalization, temporary space and memory allocation,
# and other default parameters used to determine how data is returned
# from the server to a client.
[ GENERAL ]
// Localization/Internationalization parameters.
LOCALE     =     "Italian";
SORT_ORDER_LOCALE     =     "Italian";
SORT_TYPE = "binary";
// Case sensitivity should be set to match the remote
// target database.
CASE_SENSITIVE_CHARACTER_COMPARISON = OFF ;
// SQLServer65 sorts nulls first, whereas Oracle sorts
// nulls last. This ini file property should conform to
// that of the remote target database, if there is a
// single remote database. Otherwise, choose the order
// that matches the predominant database (i.e. on the
// basis of data volume, frequency of access, sort
// performance, network bandwidth).
NULL_VALUES_SORT_FIRST = OFF;
DATE_TIME_DISPLAY_FORMAT = "yyyy/mm/dd hh:mi:ss" ;
DATE_DISPLAY_FORMAT = "yyyy/mm/dd" ;
TIME_DISPLAY_FORMAT = "hh:mi:ss" ;
// Temporary space, memory, and resource allocation
// parameters.
// You may use KB, MB for memory size.
WORK_DIRECTORY_PATHS     =     "C:\OracleBIData\tmp";
SORT_MEMORY_SIZE = 4 MB ;
SORT_BUFFER_INCREMENT_SIZE = 256 KB ;
VIRTUAL_TABLE_PAGE_SIZE = 128 KB ;
// Analytics Server will return all month and day names as three
// letter abbreviations (e.g., "Jan", "Feb", "Sat", "Sun").
// To use complete names, set the following values to YES.
USE_LONG_MONTH_NAMES = NO;
USE_LONG_DAY_NAMES = NO;
UPPERCASE_USERNAME_FOR_INITBLOCK = NO ; // default is no
// Aggregate Persistence defaults
// The prefix must be between 1 and 8 characters long
// and should not have any special characters ('_' is allowed).
AGGREGATE_PREFIX = "SA_" ;
# Security Section
# Legal value for DEFAULT_PRIVILEGES are:
# NONE READ
[ SECURITY ]
DEFAULT_PRIVILEGES = READ;
PROJECT_INACCESSIBLE_COLUMN_AS_NULL     =     NO;
MINIMUM_PASSWORD_LENGTH     =     0;
#IGNORE_LDAP_PWD_EXPIRY_WARNING = NO; // default is no.
#SSL=NO;
#SSL_CERTIFICATE_FILE="servercert.pem";
#SSL_PRIVATE_KEY_FILE="serverkey.pem";
#SSL_PK_PASSPHRASE_FILE="serverpwd.txt";
#SSL_PK_PASSPHRASE_PROGRAM="sitepwd.exe";
#SSL_VERIFY_PEER=NO;
#SSL_CA_CERTIFICATE_DIR="CACertDIR";
#SSL_CA_CERTIFICATE_FILE="CACertFile";
#SSL_TRUSTED_PEER_DNS="";
#SSL_CERT_VERIFICATION_DEPTH=9;
#SSL_CIPHER_LIST="";
# There are 3 types of authentication. The default is NQS
# You can select only one of them
#----- 1 -----
#AUTHENTICATION_TYPE = NQS; // optional and default
#----- 2 -----
#AUTHENTICATION_TYPE = DATABASE;
# [ DATABASE ]
# DATABASE = "some_data_base";
#----- 3 -----
#AUTHENTICATION_TYPE = BYPASS_NQS;
# Server Section
[ SERVER ]
SERVER_NAME = Oracle_BI_Server ;
READ_ONLY_MODE = NO;     // default is "NO". That is, repositories can be edited online.
MAX_SESSION_LIMIT = 2000 ;
MAX_REQUEST_PER_SESSION_LIMIT = 500 ;
SERVER_THREAD_RANGE = 40-100;
SERVER_THREAD_STACK_SIZE = 0; // default is 256 KB, 0 for default
DB_GATEWAY_THREAD_RANGE = 40-200;
DB_GATEWAY_THREAD_STACK_SIZE = 0; // default is 256 KB, 0 for default
MAX_EXPANDED_SUBQUERY_PREDICATES = 8192; // default is 8192
MAX_QUERY_PLAN_CACHE_ENTRIES = 1024; // default is 1024
MAX_DRILLDOWN_INFO_CACHE_ENTRIES = 1024; // default is 1024
MAX_DRILLDOWN_QUERY_CACHE_ENTRIES = 1024; // default is 1024
INIT_BLOCK_CACHE_ENTRIES = 20; // default is 20
CLIENT_MGMT_THREADS_MAX = 5; // default is 5
# The port number specified with RPC_SERVICE_OR_PORT will NOT be considered if
# a port number is specified in SERVER_HOSTNAME_OR_IP_ADDRESSES.
RPC_SERVICE_OR_PORT = 9703; // default is 9703
# If port is not specified with a host name or IP in the following option, the port
# number specified at RPC_SERVICE_OR_PORT will be considered.
# When port number is specified, it will override the one specified with
# RPC_SERVICE_OR_PORT.
SERVER_HOSTNAME_OR_IP_ADDRESSES = "ALLNICS"; # Example: "hostname" or "hostname":port
# or "IP1","IP2":port or
# "hostname":port,"IP":port2.
# Note: When this option is active,
# CLUSTER_PARTICIPANT should be set to NO.
ENABLE_DB_HINTS = YES; // default is yes
PREVENT_DIVIDE_BY_ZERO = YES;
CLUSTER_PARTICIPANT = NO; # If this is set to "YES", comment out
# SERVER_HOSTNAME_OR_IP_ADDRESSES. No specific NIC support
# for the cluster participant yet.
// Following required if CLUSTER_PARTICIPANT = YES
#REPOSITORY_PUBLISHING_DIRECTORY = "<dirname>";
#REQUIRE_PUBLISHING_DIRECTORY = YES; // Don't join cluster if directory not accessible
DISCONNECTED = NO;
AUTOMATIC_RESTART = YES;
# Dynamic Library Section
# The dynamic libraries specified in this section
# are categorized by the CLI they support.
[ DB_DYNAMIC_LIBRARY ]
ODBC200 = nqsdbgatewayodbc;
ODBC350 = nqsdbgatewayodbc35;
OCI7 = nqsdbgatewayoci7;
OCI8 = nqsdbgatewayoci8;
OCI8i = nqsdbgatewayoci8i;
OCI10g = nqsdbgatewayoci10g;
DB2CLI = nqsdbgatewaydb2cli;
DB2CLI35 = nqsdbgatewaydb2cli35;
NQSXML = nqsdbgatewayxml;
XMLA = nqsdbgatewayxmla;
ESSBASE = nqsdbgatewayessbasecapi;
# User Log Section
# The user log NQQuery.log is kept in the server\log directory. It logs
# activity about queries when enabled for a user. Entries can be
# viewed using a text editor or the nQLogViewer executable.
[ USER_LOG ]
USER_LOG_FILE_SIZE = 10 MB; // default size
CODE_PAGE = "UTF8"; // ANSI, UTF8, 1252, etc.
# Usage Tracking Section
# Collect usage statistics on each logical query submitted to the
# server.
[ USAGE_TRACKING ]
ENABLE = NO;
//==============================================================================
// Parameters used for writing data to a flat file (i.e. DIRECT_INSERT = NO).
STORAGE_DIRECTORY = "<full directory path>";
CHECKPOINT_INTERVAL_MINUTES = 5;
FILE_ROLLOVER_INTERVAL_MINUTES = 30;
CODE_PAGE = "ANSI"; // ANSI, UTF8, 1252, etc.
//==============================================================================
DIRECT_INSERT = YES;
//==============================================================================
// Parameters used for inserting data into a table (i.e. DIRECT_INSERT = YES).
PHYSICAL_TABLE_NAME = "<Database>"."<Catalog>"."<Schema>"."<Table>" ; // Or "<Database>"."<Schema>"."<Table>" ;
CONNECTION_POOL = "<Database>"."<Connection Pool>" ;
BUFFER_SIZE = 10 MB ;
BUFFER_TIME_LIMIT_SECONDS = 5 ;
NUM_INSERT_THREADS = 5 ;
MAX_INSERTS_PER_TRANSACTION = 1 ;
//==============================================================================
# Query Optimization Flags
[ OPTIMIZATION_FLAGS ]
STRONG_DATETIME_TYPE_CHECKING = ON ;
# CubeViews Section
[ CUBE_VIEWS ]
DISTINCT_COUNT_SUPPORTED = NO ;
STATISTICAL_FUNCTIONS_SUPPORTED = NO ;
USE_SCHEMA_NAME = YES ;
USE_SCHEMA_NAME_FROM_RPD = YES ;
DEFAULT_SCHEMA_NAME = "ORACLE";
CUBE_VIEWS_SCHEMA_NAME = "ORACLE";
LOG_FAILURES = YES ;
LOG_SUCCESS = NO ;
LOG_FILE_NAME     =     "C:\OracleBI\server\Log\CubeViews.Log";
# MDX Member Name Cache Section
# Cache subsystem for mapping between unique name and caption of
# members for all SAP/BW cubes in the repository.
[ MDX_MEMBER_CACHE ]
// The entry to indicate if the feature is enabled or not, by default it is NO since this only applies to SAP/BW cubes
ENABLE = NO ;
// The path to the location where cache will be persisted, only applied to a single location,
// the number at the end indicates the capacity of the storage. When the feature is enabled,
// administrator needs to replace the "<full directory path>" with a valid path,
// e.g. DATA_STORAGE_PATH = "C:\OracleBI\server\Data\Temp\Cache" 500 MB ;
DATA_STORAGE_PATH     =     "C:\OracleBIData\cache" 500 MB;
// Maximum disk space allowed for each user;
MAX_SIZE_PER_USER = 100 MB ;
// Maximum number of members in a level will be able to be persisted to disk
MAX_MEMBER_PER_LEVEL = 1000 ;
// Maximum size for each individual cache entry size
MAX_CACHE_SIZE = 100 MB ;
# Oracle Dimension Export Section
[ ORA_DIM_EXPORT ]
USE_SCHEMA_NAME_FROM_RPD = YES ; # NO
DEFAULT_SCHEMA_NAME = "ORACLE";
ORA_DIM_SCHEMA_NAME = "ORACLE";
LOGGING = ON ; # OFF, DEBUG
LOG_FILE_NAME     =     "C:\OracleBI\server\Log\OraDimExp.Log";