DEV Community: Michael

GBase 8a date_part and extract: Extracting Date/Time Components

Michael — Mon, 27 Jul 2026 15:45:00 +0000

GBase 8a provides two functions for extracting specific parts of a date or time value: date_part and extract. They are equivalent in functionality — date_part has a simpler syntax, while extract follows the Oracle‑compatible style. Both support a wide range of extraction types, from year and month down to microseconds.

Syntax

-- date_part: type can be quoted or unquoted
date_part(type, date)

-- extract
extract(type FROM date)

type: the component to extract, e.g. year, month, day.
date: a datetime expression (usually DATETIME; DATE values are auto‑cast).

Quick Reference of Common Types

type	Description	Example Output
year	Year	2023
month	Month	4
day	Day	23
hour	Hour	11
minute	Minute	12
second	Second	13
quarter	Quarter (1–4)	2
week	Week of year	17
dow	Day of week (Sunday=0)	0
isodow	ISO day of week (Monday=1, Sunday=7)	7
doy	Day of year	113
epoch	Unix timestamp (with microseconds)	1682219533.56789
year_month	Year and month	202304
day_hour	Day and hour	2311
hour_minute	Hour and minute	1112
minute_second	Minute and second	1213
millisecond	Millisecond (integer)	567
microsecond	Microsecond (integer)	567890
century	Century	21
decade	Decade (year ÷ 10)	202
millennium	Millennium	3
milliseconds	Seconds + milliseconds (decimal)	13567.89
microseconds	Seconds + microseconds (integer)	13567890

Typical Examples

All examples use date_part; extract produces identical results.

-- Year
SELECT date_part(year, '2023-04-23 11:12:13.56789');      -- 2023

-- Month
SELECT date_part(month, '2023-04-23 11:12:13.56789');     -- 4

-- Day
SELECT date_part(day, '2023-04-23 11:12:13.56789');       -- 23

-- Hour
SELECT date_part(hour, '2023-04-23 11:12:13.56789');      -- 11

-- Minute
SELECT date_part(minute, '2023-04-23 11:12:13.56789');    -- 12

-- Second
SELECT date_part(second, '2023-04-23 11:12:13.56789');    -- 13

-- Millisecond (truncates the fractional seconds)
SELECT date_part(millisecond, '2023-04-23 11:12:13.56789'); -- 567

-- Microsecond
SELECT date_part(microsecond, '2023-04-23 11:12:13.56789'); -- 567890

-- Year and month as an integer (202304)
SELECT date_part(year_month, '2023-04-23 11:12:13.56789');

-- Day and hour as an integer (2311)
SELECT date_part(day_hour, '2023-04-23 11:12:13.56789');

-- Hour and minute as an integer (1112)
SELECT date_part(hour_minute, '2023-04-23 11:12:13.56789');

-- Minute and second as an integer (1213)
SELECT date_part(minute_second, '2023-04-23 11:12:13.56789');

-- Epoch with microseconds (extract only)
SELECT extract(epoch FROM '2023-04-23 11:12:13.56789');    -- 1682219533.56789

-- Milliseconds as decimal (13.56789 seconds)
SELECT date_part(milliseconds, '2023-04-23 11:12:13.56789'); -- 13567.89

-- Microseconds as integer (13.56789 seconds → 13567890 µs)
SELECT date_part(microseconds, '2023-04-23 11:12:13.56789'); -- 13567890

-- ISO year vs. calendar year
SELECT date_part(isoyear, '2023-01-01 11:12:13');          -- 2022
SELECT date_part(year, '2023-01-01 11:12:13');             -- 2023

With these two functions, extracting any component from a date or time value becomes straightforward. They are essential for grouping, filtering, and reporting in a gbase database.

GBase 8a sha2: Built‑in SHA‑224, SHA‑256, SHA‑384, and SHA‑512 Hashing

Michael — Mon, 27 Jul 2026 15:15:00 +0000

The sha2 function in GBase 8a provides native support for multiple SHA‑2 hash algorithms. You can compute cryptographic hashes directly in SQL, which is useful for data integrity verification and secure storage in a gbase database.

Syntax

sha2(str, mode)

str: The input string. Supports Chinese, English, digits, and all characters allowed by the VARCHAR type.
mode: The hash algorithm selector. Only four values are accepted:
- 224 → SHA‑224
- 256 → SHA‑256
- 384 → SHA‑384
- 512 → SHA‑512

Any other value returns NULL. The input string should use UTF‑8 encoding.

mode	Algorithm	Return Length
224	SHA‑224	varchar(56)
256	SHA‑256	varchar(64)
384	SHA‑384	varchar(96)
512	SHA‑512	varchar(128)

Examples

-- SHA-224
gbase> SELECT sha2('GBase 8a', 224);
+----------------------------------------------------------+
| sha2('GBase 8a',224)                                     |
+----------------------------------------------------------+
| 7e3272023ef3e8edcbeb1daf2244529d806229d481753d45bd5bb524 |
+----------------------------------------------------------+

-- SHA-256
gbase> SELECT sha2('GBase 8a', 256);
+------------------------------------------------------------------+
| sha2('GBase 8a',256)                                             |
+------------------------------------------------------------------+
| 37e28220f6937e7dd7afcf6511e35a72d835dc50bd8c363fec8054485933313b |
+------------------------------------------------------------------+

-- SHA-384
gbase> SELECT sha2('GBase 8a', 384);
+--------------------------------------------------------------------------------------------------+
| sha2('GBase 8a',384)                                                                             |
+--------------------------------------------------------------------------------------------------+
| a1ea79415dc17396399a97d1e07c0553c775519be2f031333c9ef05d724341e4f2ed09cda8b890c595c1add5e8df2348 |
+--------------------------------------------------------------------------------------------------+

-- SHA-512
gbase> SELECT sha2('GBase 8a', 512);
+----------------------------------------------------------------------------------------------------------------------------------+
| sha2('GBase 8a',512)                                                                                                             |
+----------------------------------------------------------------------------------------------------------------------------------+
| 66668c66c50b095b772fc69cc903881a6107187cf33d4d9b43016c3b7260115e99175b677c2d603675ac7edcc0cf5ec7436c905d1b126237ab5e7949e48b62ed |
+----------------------------------------------------------------------------------------------------------------------------------+

-- Invalid mode
gbase> SELECT sha2('GBase 8a', 123);
+----------------------+
| sha2('GBase 8a',123) |
+----------------------+
| NULL                 |
+----------------------+

The sha2 function brings multiple SHA‑2 variants directly into your SQL toolkit. Whether you're hashing passwords or verifying row‑level integrity, you can do it inline without leaving the gbase database engine.

GBase 8a Pivot: Turning Rows into Columns

Michael — Mon, 27 Jul 2026 14:45:00 +0000

GBase 8a provides the pivot function for data pivoting — converting multiple rows into columns. This article demonstrates the syntax and common patterns using student score data in a gbase database.

Syntax

PIVOT(
  <aggregate function>(<column>)[ AS alias] [, aggregate(...)[ AS alias]...]
  FOR <col1 [, col2 ...]>
    IN (constant1 [AS alias1] [, constant2 [AS alias2] ...])
)

Aggregate part – defines the values for the new columns. Multiple aggregates are allowed. Provide aliases to avoid duplicate column names.
FOR clause – the column(s) to pivot. Single columns can omit parentheses; multiple columns must be wrapped in parentheses and the combined values treated as a single constant.
IN clause – each constant generates a new column. Wrap string/date constants in single quotes. Use aliases for readable column names, especially with multi‑column pivots.

Columns not referenced in FOR or in an aggregate function are implicitly treated as GROUP BY columns.

Sample Table and Data

CREATE TABLE tp (name VARCHAR(10), class VARCHAR(10), score INT);
INSERT INTO tp VALUES
('Zhang San','Math',99), ('Zhang San','Math',129),
('Li Si','Math',88), ('Li Si','Math',128),
('Wang Wu','Math',111), ('Wang Wu','Math',161),
('Zhang San','Chinese',120), ('Zhang San','Chinese',120),
('Li Si','Chinese',140), ('Li Si','Chinese',150),
('Wang Wu','Chinese',100), ('Wang Wu','Chinese',120);

Single Aggregate

Total score per subject:

SELECT * FROM tp PIVOT(SUM(score) FOR class IN ('Chinese','Math'));
+--------+----------+------+
| name   | Chinese  | Math |
+--------+----------+------+
| Li Si  |      290 |  216 |
| Zhang San |   240 |  228 |
| Wang Wu|      220 |  272 |
+--------+----------+------+

Multiple Aggregates

Sum and average with aliases:

SELECT * FROM tp PIVOT(SUM(score) sum, AVG(score) avg FOR class IN ('Chinese','Math'));
+--------+-------------+----------+-------------+----------+
| name   | Chinese_sum | Math_sum | Chinese_avg | Math_avg |
+--------+-------------+----------+-------------+----------+
| Li Si  |         290 |      216 |    145.0000 |  108.0000|
| Zhang San |      240 |      228 |    120.0000 |  114.0000|
| Wang Wu|         220 |      272 |    110.0000 |  136.0000|
+--------+-------------+----------+-------------+----------+

Aliasing the Constant Values

Rename the subject columns for brevity:

SELECT * FROM tp PIVOT(SUM(score) sum, AVG(score) avg FOR class IN ('Chinese' yw, 'Math' sx));
+--------+--------+--------+---------+---------+
| name   | yw_sum | sx_sum | yw_avg  | sx_avg  |
+--------+--------+--------+---------+---------+
| Li Si  |    290 |    216 | 145.000 | 108.000 |
| Zhang San|  240 |    228 | 120.000 | 114.000 |
| Wang Wu|    220 |    272 | 110.000 | 136.000 |
+--------+--------+--------+---------+---------+

Pivoting on Multiple Columns

After adding an age column, you can pivot on a combined (name, age) key. Always use aliases for readability.

SELECT * FROM tp2 PIVOT(SUM(score) sum FOR (name, age) IN (('Zhang San',40) zs40, ('Li Si',45) ls45));
+--------+----------+----------+
| class  | zs40_sum | ls45_sum |
+--------+----------+----------+
| Chinese|      240 |      290 |
| Math   |      228 |      216 |
+--------+----------+----------+

Without aliases, the column names become unreadable strings.

pivot in GBase 8a gives you native SQL support for reshaping data, making cross‑tab reports and summary tables much easier to write. Combine it with unpivot to move seamlessly between wide and long formats in your gbase database.

Fixing Rebalance Stuck During GBase 8a Mirror Cluster Expansion

Michael — Mon, 27 Jul 2026 14:11:37 +0000

When you scale out or scale in a GBase 8a mirror cluster, the operation changes the number of primary shards (hashmaps) or the distribution count. This immediately breaks the mirror prerequisite: both sides must have the same number of primary shards and compatible distributions. As a result, the rebalance process gets stuck indefinitely and cannot finish. Here is how to resolve it in your gbase database.

The Symptom

The gcluster express.log shows an error like this:

<RebalanceRunQuery|259>:gcluster command error: vc00001's hashmap number(2) do not equal vc00002's hashmap number(1)

The rebalance operation simply cannot proceed because the mirror relationship is now invalid.

The Fix

Remove the mirror relationship first. Once the expansion is complete and the hashmaps are aligned again, you can re‑establish the mirror.

ALTER DATABASE testdb DELETE MIRROR;

After running this command, the rebalance will continue and finish normally. This is the standard approach whenever a topology change invalidates a mirror configuration in GBASE's MPP cluster. Always verify that both sides have matching hashmap counts before re‑creating the mirror to keep your gbase database synchronisation healthy.

Accessing an IPv4 GBase 8a Cluster from IPv6 via the Gateway and dblink

Michael — Sun, 26 Jul 2026 15:19:00 +0000

During the transition from IPv4 to IPv6, you may have two GBase 8a clusters running on different IP protocols. Since IPv4 and IPv6 cannot communicate directly, the GBase Gateway acts as a bridge. This guide walks through the complete configuration, from gateway setup to dblink usage, for accessing an IPv4 cluster from an IPv6 cluster in a gbase database environment.

Environment

Gateway: Version GBase8a_MPP_Cluster_Gateway_8.5.1.2_build4.19, must be installed on the IPv6 cluster node (which also has an IPv4 address).
IPv4 cluster: Legacy cluster, e.g., 10.0.2.201.
IPv6 cluster: Migrated cluster with IPv6 address 2001::203, also retaining 10.0.2.203 for gateway communication.

Gateway Configuration

1. Data Source (IPv4 Cluster)

Create a file under conf/dataSource/ — the filename is the identifier you'll use in CREATE DATABASE LINK. Example ipv4_201.properties:

[ipv6201]
dataSource_IP=10.0.2.201
dataSource_port=5258
dataSource_dbname=testdb
dataSource_dbtype=gcluster
dataSource_user=gbase
dataSource_pwd=XXXXXXXX

2. Target Cluster (IPv6)

Edit conf/gcluster/gbase8a_gcluster.properties. The user must match the logged‑in user, and you must configure entries for both the coordinator (port 5258) and every data node (port 5050).

[gc1]
gcluster_IP=2001::203
gcluster_port=5258
gcluster_user=root
gcluster_pwd=XXXXXX
gcluster_encode=utf-8

[gn1]
gcluster_IP=2001::222
gcluster_port=5050
gcluster_user=root
gcluster_pwd=XXXXXX
gcluster_encode=utf-8

3. Start the Gateway

cd /opt/GBase8a_MPP_Cluster_Gateway_8.5.1.2_build4.19/
sh gt.sh

Verify with ps -ef | grep ga. Logs are in the logs directory.

IPv6 Cluster Parameters

Add the following to gbase_8a_gcluster.cnf on every coordinator node and restart:

gcluster_dblink_direct_data_exchange=0
gbase_dblink_gateway_ip=10.0.2.203
gbase_dblink_gateway_port=9898
_t_gcluster_dblink_insert_select_optimization=0

gbase_dblink_gateway_ip is the IPv4 address of the gateway host.
Setting gcluster_dblink_direct_data_exchange=0 forces all data through the gateway, which is required when the two clusters cannot reach each other directly.

dblink Usage Examples

Create a dblink

CREATE DATABASE LINK ipv4201 CONNECT TO '' IDENTIFIED BY '' USING 'ipv4_201';

Query remote data

SELECT * FROM d@ipv4201;

Import data via INSERT SELECT

INSERT INTO d SELECT * FROM d@ipv4201;

Create a table from a remote query

CREATE TABLE d2 AS SELECT * FROM d@ipv4201;

JOINs between dblink tables

JOINs between two dblink tables are supported. JOINs between a dblink table and a local table are not allowed.

Send DDL/DML to the remote cluster

Use the PASSTHROUGH syntax to execute statements directly on the IPv4 cluster:

PASSTHROUGH LINK ipv4201 USING 'INSERT INTO d VALUES(now())';

Note: SELECT and CALL are not supported via PASSTHROUGH.

View and drop dblinks

SELECT * FROM gbase.db_links;
DROP DATABASE LINK ipv4201;

Critical Notes

Install the gateway on an IPv6 cluster node — it must have both IPv4 and IPv6 addresses to bridge the two networks.
Use a consistent IPv6 address format — 2001::200 and 2001::0200 are treated as different strings. Standardise the format across the cluster.
The target cluster user in the gateway config must match the logged‑in user — matching is done by IP + username.
Configure data node entries in the gateway — if using gateway relay mode, every data node needs a gcluster_port=5050 entry.
Direct INSERT INTO table@dblink is not yet supported in the tested version.

With this setup, you can safely access an IPv4 GBase 8a cluster from an IPv6 cluster during your network migration, keeping cross‑cluster queries operational in your gbase database environment.

Checking Whether a View's Underlying Tables Still Exist in GBase 8a

Michael — Sun, 26 Jul 2026 14:16:00 +0000

When you create a view in GBase 8a, all referenced tables or other views must exist. However, if those underlying objects are dropped later, querying the view will fail. GBase 8a provides the CHECK TABLE command to detect whether a view's references are still valid, making dependency troubleshooting much easier in a gbase database.

Syntax

CHECK TABLE [vc.][db.]VIEW_NAME

Example

We'll create a base table a, a view v_a that depends on it, and then see what CHECK TABLE reports before and after the table is dropped.

1. Create the Table and View

gbase> CREATE TABLE a(id INT);
Query OK, 0 rows affected

gbase> INSERT INTO a VALUES(8);
Query OK, 1 row affected

gbase> CREATE VIEW v_a AS SELECT * FROM a;
Query OK, 0 rows affected

gbase> SELECT * FROM v_a;
+------+
| id   |
+------+
|    8 |
+------+

2. CHECK TABLE While the Base Table Exists

gbase> CHECK TABLE v_a;
+----------+-------+----------+----------------------------------------------------------------------------------+
| Table    | Op    | Msg_type | Msg_text                                                                         |
+----------+-------+----------+----------------------------------------------------------------------------------+
| test.v_a | check | Error    | (GBA-02SC-1001) The query includes syntax that is not supported by the gcluster. |
| test.v_a | check | status   | OK                                                                               |
+----------+-------+----------+----------------------------------------------------------------------------------+

The view itself is structurally sound; the status column reports OK.

3. Drop the Base Table and Check Again

gbase> DROP TABLE a;
Query OK, 0 rows affected

gbase> CHECK TABLE v_a;
+----------+-------+----------+----------------------------------------------------------------------------------------------------------------------------+
| Table    | Op    | Msg_type | Msg_text                                                                                                                   |
+----------+-------+----------+----------------------------------------------------------------------------------------------------------------------------+
| test.v_a | check | Error    | Table 'vc1.test.a' doesn't exist                                                                                           |
| test.v_a | check | Error    | View 'test.v_a' references invalid table(s) or column(s) or function(s) or definer/invoker of view lack rights to use them |
| test.v_a | check | error    | Corrupt                                                                                                                    |
+----------+-------+----------+----------------------------------------------------------------------------------------------------------------------------+

CHECK TABLE now explicitly names the missing table and marks the view as Corrupt.

When to Use This

Run CHECK TABLE whenever a view query suddenly fails with a cryptic error. It will quickly tell you whether the root cause is a missing underlying object, saving you from manually tracing dependencies across your schema. This is an essential tool for maintaining view integrity in any gbase database environment.

GBase 8a: DateDiff and TimestampDiff Have Opposite Parameter Orders

Michael — Sun, 26 Jul 2026 13:09:00 +0000

Both DATEDIFF and TIMESTAMPDIFF in GBase 8a calculate the difference between two dates, but their parameter order is reversed. This can easily lead to wrong results if you're not careful. PERIOD_DIFF follows a similar pattern. Let's clarify the differences with examples in a gbase database.

The Key Difference

DATEDIFF: Computes date1 - date2 in days. If date1 is later than date2, the result is positive.
TIMESTAMPDIFF: Computes date2 - date1 and allows you to specify the unit (year, month, day, etc.). If date1 is later than date2, the result is negative.
PERIOD_DIFF: Computes date1 - date2 in months. Arguments are integers in YYYYMM or YYMM format.

Examples

-- DATEDIFF: now() - '2004-05-01'
gbase> SELECT DATEDIFF(NOW(), '2004-05-01');
+------------------------------+
| DATEDIFF(NOW(), '2004-05-01') |
+------------------------------+
|                         6977 |
+------------------------------+

-- TIMESTAMPDIFF: '2004-05-01' - now(), returns the opposite sign
gbase> SELECT TIMESTAMPDIFF(DAY, NOW(), '2004-05-01');
+---------------------------------------+
| TIMESTAMPDIFF(DAY, NOW(), '2004-05-01') |
+---------------------------------------+
|                                 -6977 |
+---------------------------------------+

-- PERIOD_DIFF: 20230608 - 20230607
gbase> SELECT PERIOD_DIFF(20230608, 20230607);
+--------------------------------+
| PERIOD_DIFF(20230608, 20230607) |
+--------------------------------+
|                              1 |
+--------------------------------+

With the same pair of dates, DATEDIFF returns 6977 while TIMESTAMPDIFF returns -6977 — exactly the opposite sign. The reason is simply the reversed order of operands.

How to Remember

If you prefer "recent date minus old date" giving a positive number, use DATEDIFF.
When using TIMESTAMPDIFF, keep in mind it's "second argument minus first argument".
PERIOD_DIFF follows the same order as DATEDIFF.

Always double‑check which function you're using and the position of the arguments to avoid sign errors in your time‑based calculations. This small detail can prevent significant data accuracy issues in your GBASE analytical queries.

Troubleshooting gcluster_rsynctool IP Mapping Errors in GBase 8a

Michael — Sun, 26 Jul 2026 12:06:14 +0000

The gcluster_rsynctool utility synchronises tables between two GBase 8a clusters. When the clusters use private internal IPs, an IP mapping file is used to translate them into routable external IPs during synchronisation. This article covers two common connection errors related to IP mapping and how to resolve them quickly in a gbase database environment.

Error 1: `Can't connect to GBase on 'X.X.X.X:5050' (timed out)`

Symptom

The sync tool fails to connect to a cluster node's internal IP address and times out.

Root Cause

The IP mapping has not taken effect. The tool is still trying to reach the node using its private, unreachable IP instead of the translated external IP.

How to Fix

Verify that the mapping file parameters are correctly specified when launching the tool:
- --master_mpp_ip_mapping_file for the source cluster
- --slave_mpp_ip_mapping_file for the target cluster
Open the mapping file and confirm that the IP addresses are entered correctly and that the FROM (private IP) and TO (external IP) columns are not swapped.

Error 2: `Initialize master gcluster topology failed: Y.Y.Y.Y is not in mapping`

Symptom

The tool fails during topology initialisation, reporting that a specific IP Y.Y.Y.Y is not found in the mapping file.

Root Cause

A cluster node's private IP is missing from the FROM column of the mapping file, so the tool cannot translate it.

How to Fix

Check for typos in the IP addresses — the reported IP must match exactly with the entry in the mapping file.
Check the mapping direction — the format is FROM (private IP) TO (external IP). If the external IP is mistakenly placed in the FROM column, all mappings will be incorrect.

A correct mapping entry looks like this:

192.168.1.10 10.0.0.10

Here 192.168.1.10 is the internal IP reachable within the cluster, and 10.0.0.10 is the external IP visible to the other cluster.

Summary

Proper IP mapping configuration is the key to successful cross‑network synchronisation with gcluster_rsynctool. When you encounter connection timeouts or topology initialisation failures, always inspect the mapping file parameters and the file content first — this resolves the majority of such issues in a gbase database setup.

GCDW K8s 9.8.0.4.1 Multi‑Tenant Edition Overview

Michael — Wed, 22 Jul 2026 14:36:00 +0000

The GCDW (GBase Cloud Data Warehouse) 9.8.0.4.1 release introduces multi‑tenant support. Compared to the previous 9.8.0.3.5 version, the configuration has changed significantly: each tenant (account) now runs in its own dedicated Kubernetes Namespace with its own management service (gcware), coordinator, and compute warehouse.

Key Changes in values.yaml

1. Multi‑Tenant Mode

The new multiTenantMode section enables tenant isolation. When enabled, each account registered in the cloud portal becomes a tenant with an independent Namespace and dedicated resources.

multiTenantMode:
  enabled: true
  imageRegistry:
    server: "172.16.3.249:8443"
    username: "admin"
    password: "XXXXX"

In a multi‑tenant cluster, platform services (operator, redis, ftp, etc.) run in the default gcdw Namespace, while tenants such as gcdw401 and gcdwb5 have their own isolated Pods, StatefulSets, and Services.

2. FoundationDB Metadata

The metadata configuration has been simplified to a single fdb_cluster string, replacing the previously scattered FDB connection parameters.

metadata:
  fdb_cluster: "QdBWBJdf:S6QntDRlFdhyBstGIZnnQd6vVjf0ytAB@10.0.2.81:4500,10.0.2.82:4500,10.0.2.83:4500"

3. Kerberos Authentication

A Kerberos toggle is now available, indicating that GCDW supports Kerberos authentication mode (refer to the official documentation for detailed setup).

kerberos:
  enabled: true

4. Namespace Labeling

The namespaceLabel parameter allows custom labels to be added to tenant Namespaces, helping avoid conflicts with existing Namespaces.

# namespaceLabel: "Organization=gcdw"

Complete Configuration Example

Below is a full values.yaml for a multi‑tenant deployment. CPU, memory, and replica counts have been set to the minimum recommended values for production.

global:
  image:
    registry: "172.16.3.249:8443"
    pullPolicy: Always
    imagePullSecrets: "gcdw-regcred"

  multiTenantMode:
    enabled: true
    imageRegistry:
      server: "172.16.3.249:8443"
      username: "admin"
      password: "XXXXXX"

  metadata:
    fdb_cluster: "QdBWBJdf:S6QntDRlFdhyBstGIZnnQd6vVjf0ytAB@10.0.2.81:4500,10.0.2.82:4500,10.0.2.83:4500"

  storage:
    enabled: false
    storageClass: ""
    storageSize: ""

  logging:
    enabled: false
    filebeatImage: ""
    filebeatOutput:
      output.elasticsearch:
        hosts:
        - ""
        username: ""
        password: ""
        index: 'gcdw-${MY_POD_NAMESPACE}-${MY_POD_NAME}-log-%{+yyyy.MM.dd}'
    logrotateImage: "gcdw/gcdw-logrotate"

  nodeAffinitySchedule:
    cloudServiceAffinity:
      key:  cloudService
      value: gcdw
    coordiantorServiceAffinity:
      key: coordinatorService
      value: gcdw
    warehouseServiceAffinity:
      key: warehouseService
      value: gcdw

  kerberos:
    enabled: true

  # namespaceLabel: "Organization=gcdw"

operator:
  enabled: true
  image:
    repository: gcdw/gcdw-operator
    tag: 9.8.0.4.1
  replicas: 1
  leaderElect: true
  resources:
    requests:
      cpu: 1
      memory: 1Gi
    limits:
      cpu: 2
      memory: 2Gi

server:
  image:
    repository: gcdw/gcdw-server
    tag: 9.8.0.4.1
  gcware_replicas: 1
  coordinator_replicas: 1
  gcware_resource:
      request:
        cpu: "1"
        mem: "1Gi"
      limit:
        cpu: "4"
        mem: 4Gi
  coordinator_resource:
      request:
        cpu: "1"
        mem: 1Gi
      limit:
        cpu: "4"
        mem: 16Gi
  warehouse_resource:
      request:
        cpu: "1"
        mem: 1Gi
      limit:
        cpu: "16"
        mem: 32Gi

saas:
  enabled: true
  TZ: "Asia/Shanghai"
  image:
    frontend:
      repository: gcdw/clound-database-frontend
      tag: 9.8.0.4.1
    backend:
      repository: gcdw/clound-database-backend
      tag: 9.8.0.4.1
    redis:
      repository: gcdw/redis
      tag: latest
    vsftpd:
      repository: gcdw/vsftpd
      tag: latest
  replicas:
    frontend: 1
    backend:  1
    redis: 1
    vsftpd: 1
  webServicePort: 32144

Summary

GCDW 9.8.0.4.1 delivers hard resource isolation between business lines through its multi‑tenant architecture, with each tenant owning a dedicated Namespace and database service instances. When upgrading, pay close attention to the multiTenantMode and FDB metadata configuration changes, and configure Kerberos and Namespace labels as required by your environment.

In a gbase database ecosystem, this multi‑tenant GCDW release is a significant step toward true cloud‑native data warehousing, allowing multiple teams to share a single K8s cluster with full data and operational isolation.

GBase 8a unpivot Function: Turning Columns into Rows

Michael — Wed, 22 Jul 2026 13:32:00 +0000

unpivot is the counterpart of the pivot function in GBase 8a. It transforms multiple columns into rows — a classic "wide‑to‑long" transformation that is essential for normalising denormalised data. This article demonstrates the syntax and provides progressive examples using a student score table in a gbase database.

Syntax

unpivot 
  [{include_nulls | exclude_nulls}]
  (
    COLUMN | (
      COLUMN[,COLUMN]..
    )
    FOR COLUMN | (
      COLUMN[,COLUMN]..
    )
    IN COLUMN | (
      COLUMN[,COLUMN]...
    ) [ AS LITERAL | (
        LITERAL[,LITERAL]...
      )
    ]
  )

Parameter Details

NULLS — INCLUDE_NULLS keeps rows where the source column value is NULL; EXCLUDE_NULLS (the default) filters those rows out entirely.
Output columns — The function generates two new columns: one holds the original column name (or its alias), and the other holds the corresponding value.

Sample Data

The table tpp stores each student's Chinese and math scores:

gbase> select * from tpp;
+--------+--------+--------+----------+----------+
| name   | yw_sum | sx_sum | yw_avg   | sx_avg   |
+--------+--------+--------+----------+----------+
| Li Si  |    290 |    216 | 145.0000 | 108.0000 |
| Zhang San |  240 |    228 | 120.0000 | 114.0000 |
| Wang Wu |    220 |    272 | 110.0000 | 136.0000 |
+--------+--------+--------+----------+----------+

yw_sum / sx_sum — total scores for Chinese and math
yw_avg / sx_avg — average scores

Single‑Column Unpivot

Turn the yw_sum column into rows. The new column class stores the category label, and score stores the value.

select * from tpp unpivot(score for class in (yw_sum as 'Chinese Total'));

Multi‑Column Unpivot

Unpivot two columns — Chinese and math total scores — at once:

select * from tpp unpivot(score for class in (yw_sum as 'Chinese Total', sx_sum as 'Math Total'));

This produces 6 rows — two rows per student.

Mixed Data Types

Unpivot three columns, mixing integer totals and decimal averages:

select * from tpp unpivot(score for class in (yw_sum as 'Chinese Total', sx_sum as 'Math Total', yw_avg as 'Chinese Average'));

Full Unpivot of All Columns

Unpivot all four columns to obtain a clean key‑value structure:

select * from tpp unpivot(score for class in (yw_sum as 'Chinese Total', sx_sum as 'Math Total', yw_avg as 'Chinese Average', sx_avg as 'Math Average'));

Result: 12 rows, each containing the student name, the subject label, and the corresponding score.

When to Use unpivot

unpivot is especially useful when:

You receive wide‑format reports that need to be normalised for further aggregation or joining.
You need to build dynamic dashboards where each metric category should be its own row.
You are reshaping data for OLAP analysis or feeding into visualisation tools.

In a gbase database, unpivot provides a clean, SQL‑native way to flatten wide tables into a tidy long format — no complex joins or unions required. Combined with pivot, it gives you full control over data reshaping directly in your GBASE analytical queries.

Using Python with GBase 8a: Connection, CRUD, Stored Procedures, and Data Loading

Michael — Wed, 22 Jul 2026 12:17:00 +0000

This guide demonstrates how to interact with a gbase database using Python. It covers driver installation, connection setup, DDL, DML, DQL, stored procedure calls, batch inserts, and data loading. All examples use Python 2 and the official gbase-connector-python2-5.0.1 driver.

Driver Installation

Extract the tarball and install the driver:

tar xvf gbase-connector-python2-5.0.1.tar.gz
cd gbase-connector-python2-5.0.1
python setup.py build
python setup.py install

Connecting to the Database

from GBaseConnector import connect, GBaseError

config = {
    'host': '10.10.10.10',
    'port': 5258,
    'database': 'testdb',
    'user': 'root',
    'passwd': 'PASSWORD_OF_ROOT',
    'connection_timeout': 3600,
    'charset': 'utf8'
}
try:
    conn = connect()
    conn.connect(**config)
    print conn
except GBaseError.DatabaseError, err:
    print err
finally:
    conn.close()

DDL Operations

cur = conn.cursor()
cur.execute("CREATE TABLE IF NOT EXISTS testPython(id INT)")

DML Operations

cur.execute("INSERT INTO testPython VALUES(12345678)")

DQL – Querying Data

Use fetchone() or fetchall() for single queries. Always consume the entire result set before executing another query on the same cursor; otherwise, you'll get Unread result found.

cur.execute("SELECT * FROM testPython")
row = cur.fetchone()
print row  # (12345678,)

# Multiple queries
cur.execute("SELECT * FROM testPython")
rows = cur.fetchall()
cur.execute("INSERT INTO testPython VALUES(22222),(33333)")
cur.execute("SELECT * FROM testPython")
rows2 = cur.fetchmany(2)

Executing Multiple SQL Statements at Once

Set multi_stmt=True to run several statements in a single call.

cur.execute("DROP TABLE IF EXISTS testPython; CREATE TABLE IF NOT EXISTS testPython(id INT); INSERT INTO testPython VALUES(12345678)", multi_stmt=True)

# Running multiple SELECTs
rss = cur.execute("SELECT * FROM testPython; SELECT COUNT(*) FROM testPython", multi_stmt=True)
for rs in rss:
    print rs.fetchall()

Batch Insert

Use executemany() for efficient batch inserts.

opfmt = "INSERT INTO testPython2(id, name) VALUES(%s, %s)"
rows = [(i, "row" + str(i)) for i in xrange(0, 100)]
cur.executemany(opfmt, rows)

Stored Procedures

cur.execute("DROP PROCEDURE IF EXISTS testPython_proc")
cur.execute("CREATE PROCEDURE testPython_proc() BEGIN SELECT 1+2; SELECT now(); END")
rss = cur.callproc("testPython_proc")
for rs in rss:
    print rs.fetchall()
# Output: [(3,)]  [(datetime.datetime(2023, 9, 22, 9, 30, 31),)]

Loading Data

After a LOAD DATA command, inspect cur.info and cur.rowcount to obtain load statistics.

cur.execute("LOAD DATA INFILE 'file://10.10.10.10/home/gbase/py.txt' INTO TABLE testpython2 FIELDS TERMINATED BY ','")
print cur.info          # Task 3237 finished, Loaded 1 records, Skipped 0 records
print("Loaded", cur.rowcount)
infors = cur.info.split()
print("TaskId", infors[1])
print("Skipped", infors[7])

With these examples, you can quickly get started using Python with GBASE's GBase 8a for data processing and automation tasks.

Understanding GBase 8a Rebalance Logs: Stage‑by‑Stage Timing

Michael — Wed, 22 Jul 2026 11:10:07 +0000

During a scale‑out or scale‑in operation in a gbase database cluster, the rebalance process is critical for redistributing data. The cluster‑level express.log records detailed execution information, breaking the rebalance into distinct stages, each with its own timing. Analysing these stages helps you pinpoint performance bottlenecks quickly.

Sample Log

Below is a typical rebalance log entry from express.log:

Stage Descriptions

Each line in the log represents a stage, with the elapsed time shown in seconds.

Stage Name	Description
REB_MAKE_POLICY	Time spent generating the redistribution policy
REB_EXEC_POLICY	Time spent executing the redistribution policy
REB_CREATE_TABLE	Time spent creating intermediate table structures
REB_CREATE_SLICE	Time spent creating table slices
REB_SIS	Time spent executing SELECT INTO SERVER
REB_GC_SYNC	Time spent synchronising data via gc_sync_client
REB_RENAME	Time spent renaming the intermediate table to the target table
REB_CHANGE_DIST_ID	Time spent updating the distribution ID
REB_DELETE_SLICE	Time spent removing old table slices
REB_GC_LOCK	Time spent acquiring the cluster lock
REB_UPDATE_STATUS	Time spent updating progress information in system tables
REB_SIS_COMMIT	Time spent committing the SIS transaction
REB_SIS_ROLLBACK	Time spent rolling back the SIS transaction (appears only on failure)
REB_UPDATE_FAILOVER_INFO	Time spent updating failover information
REB_DATA_MOVE	Time spent performing the actual data migration

How to Interpret the Timings

Different table types show different bottlenecks:

Large tables (high row count): The dominant phases are REB_DATA_MOVE (data migration) and REB_GC_SYNC (data synchronisation).
Wide tables or tables with many partitions but few rows: The creation stages (REB_CREATE_TABLE and REB_CREATE_SLICE) may consume the most time.
Tables that complete the entire rebalance in under 10 seconds: These generally require no tuning.

By focusing on the stages with the highest elapsed times, you can target your optimisation efforts — adjusting parallelism, redesigning table structures, or expanding disk capacity where it matters most.

In a gbase database, the express.log rebalance entries are one of the most direct windows into the cluster's internal behaviour during scaling operations. Keep them handy whenever you plan a topology change.

DEV Community: Michael

GBase 8a date_part and extract: Extracting Date/Time Components

Syntax

Quick Reference of Common Types

Typical Examples

GBase 8a sha2: Built‑in SHA‑224, SHA‑256, SHA‑384, and SHA‑512 Hashing

Syntax

Examples

GBase 8a Pivot: Turning Rows into Columns

Syntax

Sample Table and Data

Single Aggregate

Multiple Aggregates

Aliasing the Constant Values

Pivoting on Multiple Columns

Fixing Rebalance Stuck During GBase 8a Mirror Cluster Expansion

The Symptom

The Fix

Accessing an IPv4 GBase 8a Cluster from IPv6 via the Gateway and dblink

Environment

Gateway Configuration

1. Data Source (IPv4 Cluster)

2. Target Cluster (IPv6)

3. Start the Gateway

IPv6 Cluster Parameters

dblink Usage Examples

Create a dblink

Query remote data

Import data via INSERT SELECT

Create a table from a remote query

JOINs between dblink tables

Send DDL/DML to the remote cluster

View and drop dblinks

Critical Notes

Checking Whether a View's Underlying Tables Still Exist in GBase 8a

Syntax

Example

1. Create the Table and View

2. CHECK TABLE While the Base Table Exists

3. Drop the Base Table and Check Again

When to Use This

GBase 8a: DateDiff and TimestampDiff Have Opposite Parameter Orders

The Key Difference

Examples

How to Remember

Troubleshooting gcluster_rsynctool IP Mapping Errors in GBase 8a

Error 1: Can't connect to GBase on 'X.X.X.X:5050' (timed out)

Symptom

Root Cause

How to Fix

Error 2: Initialize master gcluster topology failed: Y.Y.Y.Y is not in mapping

Symptom

Root Cause

How to Fix

Summary

GCDW K8s 9.8.0.4.1 Multi‑Tenant Edition Overview

Key Changes in values.yaml

1. Multi‑Tenant Mode

2. FoundationDB Metadata

3. Kerberos Authentication

4. Namespace Labeling

Complete Configuration Example

Summary

GBase 8a unpivot Function: Turning Columns into Rows

Syntax

Parameter Details

Sample Data

Single‑Column Unpivot

Multi‑Column Unpivot

Mixed Data Types

Full Unpivot of All Columns

When to Use unpivot

Using Python with GBase 8a: Connection, CRUD, Stored Procedures, and Data Loading

Driver Installation

Connecting to the Database

DDL Operations

DML Operations

DQL – Querying Data

Executing Multiple SQL Statements at Once

Batch Insert

Error 1: `Can't connect to GBase on 'X.X.X.X:5050' (timed out)`

Error 2: `Initialize master gcluster topology failed: Y.Y.Y.Y is not in mapping`