HiddenMerit Daily · Issue 43

# 📊 HiddenMerit Daily · Issue 43

> Focus on Database Frontiers, Practical Insights for DBAs

> June 15, 2026 ｜ 5 Selected Global Breaking News

## 01｜One‑Third of Central SOEs Choose OceanBase for Core Systems! OceanBase Launches “AI‑Integrated Data Solution for Central and State‑owned Enterprises”

On June 11, at the “2026 Information Technology Application Innovation Development Conference” held in Jinan, Shandong, OceanBase’s “Provincial Government Cloud Integrated Database Platform Solution” was selected as a “Typical Solution” in the “2025 Information Technology Application Innovation Solutions” category, receiving authoritative recognition from the Cybersecurity Center of the Ministry of Industry and Information Technology (MIIT). Ding Rong, General Manager of OceanBase’s Government and Enterprise Business Unit, revealed that among nearly 100 central SOEs nationwide, over 30 have entrusted their core systems to OceanBase, with hundreds of central and state‑owned enterprises using OceanBase to run key business systems, covering vital industries such as telecommunications, transportation, and energy.

Key Industry Implementation Data:

| Industry | Key Data |

|----------|----------|

| Civil Aviation | China Travelsky’s core system has been upgraded to a distributed domestic system, ensuring safe travel for over 700 million passengers annually; China Southern Airlines and Sichuan Airlines have subsequently followed suit. |

| Telecommunications | Serves one‑third of China Mobile’s provincial companies, supporting nearly 600 million users; through cooperation with China Unicom, covers over 1,000 nodes and more than 300 core systems. |

| Energy | Sinopec’s new generation smart fuel card system covers nearly 30,000 gas stations nationwide, processing 50,000 transactions per minute, achieving 8‑fold savings in storage costs. |

| Electric Power | Full‑chain coverage from power generation and transmission to consumption, stably supporting power market trading. |

| Government Affairs | Covers 22 provinces and 20 categories of government scenarios, serving the health system for 86 million people in Jiangsu and 66.7 million users of Zhejiang’s “Zheliban” platform. |

At the conference, OceanBase officially launched its “AI‑Integrated Data Solution for Central and State‑owned Enterprises.” The solution builds a unified digital foundation based on an “AI database” to support a range of AI‑driven innovation scenarios, including business management and control, supply chains, asset operations, and data intelligence, helping central and state‑owned enterprises upgrade from ERP (Enterprise Resource Planning) to DRP (Data Resource Planning) . According to a CCID Consulting report, OceanBase was rated as the “Number One Local Vendor in Development Capability and Market Position for Central and State‑owned Enterprises.”

At the conference, Ding Rong stated: “If the data foundation is unstable, the AI above it is a castle in the air. The requirements for databases in the AI era have undergone a fundamental shift – data must move from ‘being storable’ to ‘being usable, computable, and enabling fast decision‑making.’”

- DBA Perspective: Over 30 central SOEs choosing OceanBase for their core systems means that large‑scale replacement by domestic distributed databases in key sectors of the national economy has moved from “pilot projects” to “batch delivery.” DBA operations positions for domestic databases are shifting from “edge systems” to “core systems.” The real‑world data from Sinopec – 30,000 gas stations and 50,000 transactions per minute – is an important reference for DBAs evaluating OceanBase’s stability in high‑concurrency scenarios. The case of 8‑fold storage cost savings also provides a quantitative basis for DBAs conducting TCO analysis during technology selection. The concept of upgrading from ERP to DRP (Data Resource Planning) means that DBAs will no longer be just “database administrators,” but “data asset architects” – responsible for designing data circulation paths and formulating data governance rules.

- CTO Perspective: OceanBase’s penetration rate in central SOE core systems has exceeded one‑third (30+/100), confirming that the reliability of domestic distributed databases in vital industries such as telecommunications, transportation, and energy has been validated by leading users. The case of Travelsky ensuring travel for 700 million passengers provides a benchmark reference for CTOs making selections in high‑availability scenarios such as civil aviation and transportation. The “ERP → DRP” upgrade path is also worth deep consideration by CTOs when planning enterprise data strategies – the full circulation of data assets and transparent supervision are becoming new directions for enterprise digital transformation.

- Investor Perspective: OceanBase receiving the “Typical Solution” certification from the MIIT Cybersecurity Center, combined with the data of over 30 central SOEs entrusting their core systems, further solidifies its leading position in the Xinchuang market. Its horizontal expansion path from government services (covering 22 provinces) to central and state‑owned enterprises (30+ central SOEs) is clear. CCID Consulting’s “Number One Local Vendor” evaluation provides authoritative endorsement for OceanBase’s valuation in the capital market. Vendors with the capability of “AI database + full‑stack Xinchuang adaptation” will occupy stronger pricing power in the deep‑water Xinchuang replacement.

## 02｜AI Reshapes the Underlying Logic of Databases: Database Industry Evolves from “Warehouse” to “Intelligent Infrastructure”

On June 11, Xinhua Finance/Shanghai Securities News published an in‑depth report titled “AI Reshapes the Underlying Logic, Databases Ride the Wind Again.” The report points out that when enterprises no longer just ask “can we store the data,” but “can large models directly use my data to answer questions,” databases – seemingly mundane foundational software – are once again riding the wind.

Core Views:

- Agents are the New Users: Wang Yicheng, Vice President of Tencent Cloud, stated that the industry is redesigning database product capability systems with agents as the new users, and the database industry is entering the AI 3.0 era.

- Profound Change in the Database Mission: In the past, databases served programmers, BI systems, and deterministic business processes. In the future, a more important task is to enable data to be understood and used by hundreds or thousands of agents. “The way agents are developed and the way they use data have fundamentally changed – this will inevitably force databases to be restructured from the underlying architecture upward.”

- Future Software is “Agent + Database”: Zhou Aoying, Fellow of the CCF and Director of the Database Committee, pointed out that future software will be “agent + database,” and databases should become a reliable, accessible, and efficient infrastructure like the power grid.

Two Major Technology Paths in the Industry:

1. AI‑Integrated: Adding AI capabilities to traditional relational databases. Representative products include Dameng DM9, Alibaba Cloud PolarDB, OceanBase, and Tencent Cloud TDSQL.

2. AI‑Native Vector Databases: Designed from scratch, converting text and images into “digital fingerprints” for semantic retrieval. Representative products include Tencent Cloud Vector Database, Zilliz, etc.

Market Size Forecast: According to data from the “AI‑Native Database Development Trends White Paper” jointly released by IDC and Mobile Cloud, the Chinese database market is projected to reach $10.6 billion in 2026, with the domestic replacement rate exceeding 70%, shifting from “Xinchuang replacement” to “AI‑driven incremental support.”

- DBA Perspective: This report provides DBAs with a macro framework for understanding industry change. The judgement that “agents are the new users” echoes the concept of “taking agents as new users” reported in Issues 39 and 41 of our publication. For DBAs, this means that the primary access subjects of databases in the future will shift from “humans” to “AI agents,” requiring DBAs to redesign database permission governance systems, audit tracking mechanisms, and resource isolation strategies. The statement that “databases should become infrastructure like the power grid” means the DBA role will evolve from “administrator” to “infrastructure architect.”

- CTO Perspective: The IDC forecast of a $10.6 billion Chinese database market with a domestic replacement rate exceeding 70% provides CTOs with a macro‑quantitative basis for formulating medium‑ to long‑term Xinchuang plans. The comparison of the two technology paths in the report – AI‑integrated vs. AI‑native vector databases – also provides CTOs with a clear decision‑making framework for technology selection. The judgement by CCF Fellow Zhou Aoying that “future software is agent + database” is worth deep consideration by technology decision‑makers.

- Investor Perspective: The shift of databases from “Xinchuang replacement” to “AI‑driven incremental support” means that the valuation logic for domestic database vendors is upgrading from “policy‑driven” to “technology‑driven.” With the market projected to reach $10.6 billion in 2026, the growth space for leading vendors remains vast. The technology path choices of leading players such as Tencent Cloud, Dameng, and OceanBase in this AI transformation will determine their valuation differentiation in the capital market.

## 03｜PostgreSQL 19 Beta 1 Released: Graph Queries, Parallel Vacuum, REPACK, and Other Major Features Debut

On June 4, the PostgreSQL Global Development Group officially announced that PostgreSQL 19 Beta 1 is available for download. Following PG 18, PG 19 achieves comprehensive upgrades across five dimensions: performance, developer experience, security, observability, and logical replication.

Performance Enhancements:

- Automatic I/O Worker Scaling: io_method=worker now automatically adjusts the number of I/O workers based on io_min_workers and io_max_workers.

- Parallel Autovacuum: Adds the autovacuum_max_parallel_workers configuration, supporting parallel vacuum; a new scoring system helps prioritise tables needing vacuum.

- Brand‑new REPACK Command: Adds the REPACK command and a non‑blocking CONCURRENTLY option, significantly reducing the operational overhead of table rebuilds.

- Foreign Key Check Optimisation: Insert performance improves by up to 2x when foreign key constraints exist.

- New Eager Aggregation: Accelerates row processing; the optimiser adds the enable_eager_aggregate switch.

Developer Experience Enhancements:

- SQL/PGQ Property Graph Queries: First‑time support for SQL/PGQ standard, allowing users to execute property graph queries using standard SQL syntax.

- Time‑Series Query Enhancements: UPDATE and DELETE now support the FOR PORTION OF clause.

- Partition Table Reorganisation: Adds ALTER TABLE ... MERGE PARTITIONS and SPLIT PARTITIONS.

- WAIT FOR LSN: Allows a replica to wait for a specified LSN to replay before executing a query, enabling “read‑your‑writes” consistency.

- GROUP BY ALL: Simplifies grouping operations.

Security and Observability:

- SNI Support: Through the new pg_hosts.conf file, a single server can present different TLS certificates based on the hostname requested by the client.

- Password Expiry Warning: Adds password_expiration_warning_threshold (default 7 days).

- MD5 Authentication Warning: A warning is sent to the client after successful MD5 authentication.

- pg_stat_lock View: Reports statistics by lock type.

- pg_stat_recovery View: Provides detailed visibility into recovery operation status.

- New IO Option in EXPLAIN ANALYZE: Supports displaying asynchronous I/O statistics.

Logical Replication and Federated Queries:

- Sequence Value Replication: Logical replication now supports sequence value replication, simplifying online upgrade tasks.

- CREATE PUBLICATION ... EXCEPT: Publishes all tables except a specified set.

- CREATE SUBSCRIPTION ... SERVER: Uses a foreign server definition for subscriptions, simplifying credential management.

- Enable Logical Replication Without Restart: Logical replication can now be enabled without restarting the server.

- DBA Perspective: PG 19 Beta 1 is a landmark release for the PostgreSQL community. The kernelisation of the REPACK command ends the long history of DBAs relying on the pg_repack extension to solve table bloat, greatly improving stability. SQL/PGQ property graph queries mean DBAs can handle both relational data and graph queries within the PostgreSQL kernel, eliminating the need for a separate graph database. Parallelisation of autovacuum directly addresses the pain point of lengthy vacuum times on large instances. WAIT FOR LSN makes “read‑your‑writes” consistency possible in read‑write separation architectures. DBAs with PostgreSQL operations experience are advised to deploy PG 19 Beta 1 in a test environment immediately, focusing on the real‑world performance of REPACK, graph queries, and parallel vacuum. Production environments should plan upgrade windows after the final release.

- CTO Perspective: PG 19 is another important milestone in PostgreSQL’s positioning as the “data foundation for the AI era.” The introduction of property graph queries (SQL/PGQ) greatly enhances PostgreSQL’s applicability in scenarios such as enterprise knowledge graphs and intelligent recommendations. Support for logical replication of sequence values and enabling logical replication without restart reduce the complexity of online upgrades and system migrations. For enterprises with PostgreSQL as their core technology stack, these PG 19 features are worth incorporating into medium‑ and long‑term technology planning.

- Investor Perspective: PostgreSQL’s strong growth momentum (DB-Engines H1 growth ranking: +21.97 points) is validated by the features in PG 19. The kernelisation of REPACK, graph queries, and logical replication enhancements will further solidify PG’s position in the enterprise market. Commercial service companies in the PostgreSQL ecosystem (managed services, migration tools, performance optimisation products) will benefit from the upgrade demand and new application scenarios unlocked by PG 19.

## 04｜Dameng Explains AI Strategy in Detail: AI4DB and DB4AI Dual‑Drive, Global Expansion Accelerates

On June 4, Dameng hosted institutional investor research visits from 10 institutions including Eastmoney Securities, Industrial Securities, TF Securities, Taikang Asset Management, and Ping An Pension, elaborating in detail on the company’s strategic layout in the AI era.

Dual AI Strategy Tracks:

- AI4DB (AI Empowers Databases) : Artificial intelligence deeply integrates into the entire database lifecycle, leveraging AI to complete parameter auto‑tuning, intelligent index selection, SQL slow query optimisation, real‑time anomaly diagnosis, and fault self‑healing. At the same time, it intelligently schedules server CPU, memory, and I/O resources, significantly reducing manual operations pressure and achieving database auto‑management and autonomy.

- DB4AI (Databases Support AI) : Focuses on breaking through core technologies such as multi‑modal data storage and computation integration, native integrated architecture design, and database intelligent upgrades, enabling databases to natively integrate core capabilities such as vector computing, knowledge graph analysis, and intelligent analysis. Dameng’s database management system now natively supports vector data types, providing complete vector processing functionality.

Multi‑Modal Convergence Strategy: The company continues to promote unified management and collaborative processing of multiple data models – graph, relational, document, vector – effectively reducing users’ total costs in product procurement, technology learning, application development, data migration, and operations management.

The Future of Traditional Centralised Databases: Dameng judges that traditional relational databases, leveraging their advantages in transaction consistency, mature ecosystems, and data standardisation, will continue to play a key role in core business scenarios such as financial transactions, enterprise ERP, and government systems for a long time. Centralised and AI‑oriented new data management capabilities are not simply a replacement relationship, but are complementary and will coexist for a long time.

Global Expansion Layout: The company has established a wholly‑owned subsidiary, Hainan Dameng International Data Technology Co., Ltd., in Hainan. Leveraging the Hainan Free Trade Port as an international business headquarters, it continues to deepen its overseas market strategic layout, optimise product international standard compatibility, form a professional overseas operations team, and improve its overseas technical service and support system.

- DBA Perspective: Dameng’s AI strategy is clearly divided into two directions: “AI4DB” and “DB4AI.” AI4DB means that database operations are evolving from “passive firefighting” to “active self‑healing” – the “root cause reasoning + dynamic self‑healing strategy selection” mechanism in Dameng’s patent will significantly reduce DBAs’ manual troubleshooting workload. The “native vector data type” already supported in the DB4AI direction means DBAs can manage multi‑modal data within a single database. Dameng’s global expansion layout also indicates that the application scenarios for domestic database skills are expanding from “domestic replacement” to “international competition.”

- CTO Perspective: Dameng’s judgement that “centralised and AI‑oriented new data management capabilities will coexist for a long time” provides enterprises with a clear roadmap for technology selection – no need to choose between “traditional” and “AI,” but to progress in phases. The dual‑track AI strategy of AI4DB and DB4AI disclosed in Dameng’s investor research provides CTOs with a reference framework for evaluating the AI maturity of Dameng’s products.

- Investor Perspective: Dameng’s continued investment in the AI direction one year after its listing, combined with its 2025 net profit of RMB 515 million and extremely high gross margin, provides strong valuation support for it in the domestic database track. The dual‑track AI4DB and DB4AI strategy is clear, and its global expansion has already begun. Investors should focus on the commercialisation cadence of Dameng’s AI operations direction and the progress of its overseas business expansion.

## 05｜EDB Korea President: PostgreSQL is the Best Choice in the “Post‑Oracle Era,” EDB is the Only PostgreSQL Vendor in Korea

On June 15, Korean IT media Bloter reported an interview with Kim Hee‑bae, President of EDB Korea. Kim stated that PostgreSQL is becoming the most compelling alternative to Oracle databases, and that EDB, as a direct developer and modifier of PostgreSQL core code (a Vendor), is the only company in Korea with this capability.

Key Case Data:

- Shinhan EZ Insurance (core transaction system), Hyundai Motor (production system), and Kyobo Bookstore (information system) have all replaced Oracle with EDB PostgreSQL.

- Kyobo Bookstore: After 6 years of OLTP operation, total cost of ownership (TCO) savings reached 1 billion KRW (approximately RMB 5.3 million). They recently also introduced the analysis‑dedicated product “Warehouse PG.”

- India: The core system of a mobile bank used by 120 million users has adopted EDB.

EDB’s Core Advantages:

- Oracle Compatibility >95%: Simply by installing, over 95% of applications originally built on Oracle can run directly without rewriting.

- Agentic Lakehouse Strategy: EDB recently proposed the “Agentic Lakehouse” concept, interacting with various AI models through the MCP protocol to achieve 5‑stage AI autonomous operations: log collection → semantic parsing → anomaly detection → root cause reasoning → self‑healing.

- Korean Market: Already has 200 customers. Kim stated that this year’s goal is to have half of its revenue come from analytics and AI.

Kim Hee‑bae emphasised: “In the event of a failure in a sensitive system, the only company in Korea that can take immediate action is EDB.” He explained that the PostgreSQL ecosystem is divided into Vendors (core code developers) and Contributors (code reviewers and testers). Vendors are like teams that design engines and drive shafts, while Contributors are like teams that manufacture tyres. Only a Vendor can directly fix defects in the database kernel.

- DBA Perspective: EDB’s “Oracle compatibility >95%” is a core reference indicator for DBAs conducting Oracle migrations. Kyobo Bookstore’s case of saving 1 billion KRW in TCO over 6 years provides a quantitative basis for DBAs’ cost arguments during technology selection. EDB’s proposed “Agentic Lakehouse” concept – 5‑stage AI autonomous operations from anomaly detection to self‑healing – echoes the technical approaches of Dameng’s AI intelligent O&M patent and Tencent Cloud’s DatabaseClaw reported in previous issues. For DBAs, this means future operations tools will have full‑chain capability of “automatic discovery → automatic diagnosis → automatic repair.”

- CTO Perspective: EDB’s positioning as a PostgreSQL core Vendor answers the key question for CTOs when selecting open‑source databases – “Who can fix it if something goes wrong?” EDB’s scale of 200 customers in Korea, and cases in critical scenarios such as Shinhan EZ Insurance (core transactions) and Hyundai Motor (production systems), validate the reliability of PostgreSQL in enterprise core systems. If EDB’s “Oracle compatibility >95%” is verified in real‑world scenarios, it will significantly reduce the risk and workload of migrating from Oracle.

- Investor Perspective: EDB’s success in the Korean market (200 customers) and the case of a 120‑million‑user mobile bank in India demonstrate the large‑scale potential of commercial services around the PostgreSQL ecosystem. EDB’s proposed “Agentic Lakehouse” concept represents a cutting‑edge direction for AI‑driven database operations. Investors should focus on commercial service companies in the PostgreSQL ecosystem, as well as the differentiated competitive advantage of vendors with “kernel‑level repair capability” (Vendor status) in the market.

## 📚 SQL Little Knowledge Point

This Issue’s Knowledge Point: What is SQL/PGQ (Property Graph Query)?

SQL/PGQ (Property Graph Query) is an extension of SQL defined by the ISO/IEC 39075:2024 standard, allowing property graph queries to be executed directly within SQL. PostgreSQL 19 Beta 1 introduces support for this standard for the first time.

Traditional SQL vs. Graph Query:

| Dimension | Traditional SQL | SQL/PGQ Graph Query |

|-----------|-----------------|---------------------|

| Data Model | Tables, rows, columns | Vertices, edges, properties |

| Relationship Expression | Multi‑table JOINs | Path pattern matching |

| Typical Scenarios | Transaction processing, reporting | Social networks, knowledge graphs, fraud detection |

| Query Complexity | Grows exponentially with number of JOINed tables | Path depth impact on performance is relatively manageable |

Graph Query Example in PG 19 (from official documentation):

```sql

-- Create a property graph

CREATE PROPERTY GRAPH bank_graph

VERTEX TABLES (account, customer)

EDGE TABLES (

transfer AS transfer_edge

SOURCE KEY (from_account) REFERENCES account (account_id)

DESTINATION KEY (to_account) REFERENCES account (account_id)

);

-- Graph query: find money transfer paths from account A to account B

SELECT * FROM GRAPH_TABLE(bank_graph

MATCH (src IS account WHERE src.account_id = 'A')

-[trans IS transfer_edge]->+(dest IS account WHERE dest.account_id = 'B')

COLUMNS (src.account_id AS from_account,

dest.account_id AS to_account,

ARRAY_AGG(trans.amount) AS transfer_amounts));

```

Significance for DBAs:

- Knowledge graph applications can be completed within the PostgreSQL kernel, eliminating the need for an additional graph database.

- Graph queries and relational queries can be used together within the same transaction.

- Reduces data movement and consistency challenges from multi‑component integration.

PostgreSQL 19 Beta 1’s SQL/PGQ implementation marks a critical step for PostgreSQL towards becoming a “multi‑model database.”

> HiddenMerit Team Production

> Slogan: 绩优隐于内，金石启新程 | Hidden deep. Merit bold. Forge ahead.