Tuesday, December 04, 2018

Generate Serial Number for Existing Table

Table T_REVIEW has column id as the unique key and dt as the timestamp. I want to add a serial number to the table based on the order by dt and id. In the following scripts, I use window function row_number to generate the serial number and update the table.
alter table t_review add(seq number);

update t_review a set seq= (
with tbl as (select id, row_number() over(order by dt, id) rnk from t_review)
select rnk from tbl b 
where b.id=a.id
);

Incrementally Add New Records to Table

I have a table t_review that stores historical records including key sid and timestamp dt. Every day, more records come into table t_new. I use the following scripts to add those new records identified by sid in t_new to t_review.
begin
insert into t_review(sid, dt) select sid, sysdate from 
 (select sid from t_new minus select sid from t_review);
commit;
end;

Friday, November 30, 2018

Load Text Into Oracle Table CLOB Column Using Python Part 2

The texts in some records are too long, e.g., more thatn 22k characters. This is inconvenient when we try to edit it online. I use chunkstring(line2,900) to divide the text into multiple strings of fixed length of 900 characters and stored them in multiple records. uuid.uuid1().hex is used to generate an universally unique id.
import cx_Oracle
import uuid
def chunkstring(string, length):
    return (string[0+i:length+i] for i in range(0, len(string), length))
connection = cx_Oracle.connect("xxxxx","xxxx",\
                               "golden.maxapex.net:1521/xxxx",\
                               encoding = "UTF-8", nencoding = "UTF-8")
cur = connection.cursor()
cur.setinputsizes(T1 = cx_Oracle.CLOB)
cur.setinputsizes(T2 = cx_Oracle.CLOB)
f = open('id_baike2.txt',  \
        encoding="utf8")
line1 = f.readline()
line2 = f.readline()
while line2:
    vuid = uuid.uuid1().hex
    data_list = list(chunkstring(line2,900))
    for idx, v in enumerate(data_list):
        cur.execute("INSERT INTO t_text_clob_m (ID, SID, SEQ, T1)  \
               VALUES (:ID, :SID, :SEQ, :T1)",\
                    ID = int(line1), SID = vuid, SEQ = idx+1, T1 = v)
    # create table t_text_clob_m (id number, sid raw(32), seq number, t1 clob);
    line1 = f.readline()
    line2 = f.readline()
f.close()
connection.commit()
cur.close()
connection.close()

Thursday, November 29, 2018

Load Text Into Oracle Table CLOB Column Using Python

A text file contains Chinese texts that describe close to 4 thousand small towns. The ID of a town is always the first line followed by the text. The text could be as long as 22k characters. Thus, we have to use CLOB to store these texts.
1
宁明县爱店镇成立于1992年3月,位于中越陆地边境公母山下,与越南谅山省禄平县接壤,
边境线长达25.5公里,...
2
林逢镇位居田东县城东部,地处右江河谷中心地带、大西南出海通道咽喉要地,东经南宁可达北海、
防城及广东沿海各地,...
From SQL Developer client, I connect to the Oracle database and create a table with a CLOB column to store the text.
#create a table to hold texts
create table t_text (id number, val clob);
I load the file into the table on Amazone RDS Oracle database using the following scripts.
import cx_Oracle
connection = cx_Oracle.connect("xxxxx","xxx",\
                               "orcl.xxxxxx.us-east-1.rds.amazonaws.com:1521/orcl",\
                               encoding = "UTF-8", nencoding = "UTF-8")
cur = connection.cursor()
cur.setinputsizes(VAL = cx_Oracle.CLOB)
f = open('id_baike2.txt',  encoding="utf8")
line1 = f.readline()
line2 = f.readline()
while line2:
    cur.execute("INSERT INTO t_text (ID, VAL) VALUES (:ID, :VAL)", \
                ID=int(line1),VAL=line2)
    line1 = f.readline()
    line2 = f.readline()
f.close()
connection.commit()
cur.close()
connection.close()

Connect to Amazon RDS Oracle Database Using Python

The following scripts connect to an Oracle database and execute queries.
import cx_Oracle
connection = cx_Oracle.connect("aisxxx","xxxxxxx",\
              "orcl.ctbxxxxxxx.us-east-1.rds.amazonaws.com:1521/orcl")
# username, password, host:port number/service name
cur = connection.cursor()
# query table
cur.execute("select table_name, tablespace_name from user_tables order by 1")
for each in cur.description:
  print(each[0:2])
for col_1, col_2 in cur.fetchall():
 print(col_1+" "+col_2)
cur.close()
connection.close()
Outputs are the following.
('TABLE_NAME', )
('TABLESPACE_NAME', )
T_TEST USERS

Wednesday, November 28, 2018

Find All Unique Characters In a Text File

There is a file containing 7,288 lines of Chinese text (see below). I want to find our all characters that appear in the text.
$ head -2 id_baike2.txt
1
宁明县爱店镇成立于1992年3月,位于中越陆地边境公母山下,与越南谅山省禄平县接壤,
边境线长达25.5公里,东兴至那坡沿边公路贯穿其间。
爱店为国家一类口岸,是我国西南经济板块通往东南亚的陆路效能要道。
全镇下辖3个村委会19个自然屯,总面积65.79平方公里,耕地面积7225亩,
粮食播种面积2700亩,总人8055口人,其中流动人口2000余人。
爱店镇1995年被国家建设部定为全国500家小城镇建设试点之一,
2002年被定为全区小城镇建设重点镇。城镇各种基础设施日臻完善。
镇区主要街道全部实现硬化和街砖铺设。
2006年以来,爱店镇结合崇左市开展市容环境综合整治竞赛活动和城乡清洁工程,
与有关部门共同筹措资金共561.7万元投入小城镇建设,
不断扩大城镇规模和完善城镇功能,搞好环境卫生,提高口岸服务水平,
树立良好国门形象。著名的“金牛潭”风景区,
常年流水潺潺,怪石嶙峋,古木参天,景色迷人。
历代的文人名士曾在潭边题字留墨,常年吸引大批游客前来观光旅游。
镇内公母山海拔1358米,山奇水秀,景色宜人。主要有金牛潭生态旅游度假村、
公母山庄、爱店起义纪念碑、

I use the following Linux command to accomplish this.
cat  id_baike2.txt | awk -F "" '{for (i=1;i <= NF;i++) print $i;}' | sort | 
uniq -c > char_list.txt

The following shows the partial result. The number on the left is the frequency of the characters appearing in the file
      3  
  50033  
     94 !
   3773 "
     47 #
    124 $
  23974 %
     56 &
    422 '
   7333 (
   7245 )
    101 *
    672 +
   7558 ,
   9437 -
  64329 .
   1700 /
 267266 0
 239610 1
 201818 2
 108849 3
  91833 4
 111514 5
  86085 6
  69366 7
  83533 8
 100306 9
   1309 :
    758 ;
     25 <
I also use the following python scripts to do the same calculation.
import collections
f = open('id_baike2.txt',  encoding="utf8")
lines = f.readlines()
lines_joined =  ' '.join(lines)
frq = collections.Counter(lines_joined)
frq.most_common(100)
f.close()

Friday, July 27, 2018

About Dr. Zhou's Oracle SQL for Data Science Course

On January 31, 2017, I was invited by Prof. Wei Ding at the Department of Computer Science, University of Massachusetts Boston, and gave 3 talks about my data science projects across different industries. These talks are extremely well received. The following is what Prof. Ding says about my talks.

"It was a fortune to have Jay come to our computer science department to share his experience in solving business problems with predictive analytics on February 28, 2017. What Jay had presented in his 3 talks, each lasting for 1 hour in different topics of data mining, was totally impressive and beyond our wildest expectation. Having built competition-winning predictive models for some of the biggest companies and produced hundreds of millions of dollars’ savings, Jay shared the secret of his success with students and faculty without reservation. His strong presentations were such an inspiration for our computer science students and faculty and his methodology was innovative and powerful, even for very seasoned data scientists among the audience. Jay, thank you so much for your hard work preparing and delivering these presentations!" -Prof. Ding Wei, Department of Computer Science, University of Massachusetts Boston

The audience are particularly amazed by how I come up with solutions using Oracle SQL environment. To share my expertise, I create the online course Oracle SQL for Data Science to show how to perform common data science tasks using Oracle SQL and the benefits for doing that. The tuition for the course is $14.95.

I let Charlie Berger,Senior Director of Product Management, Machine Learning, AI and Cognitive Analytics at Oracle know about my course and he told me "Your course is amazing."

Thursday, June 28, 2018

Introducing New Book: Principles of Database Management – The Practical Guide to Storing, Managing and Analyzing Big and Small Data


Deep-Data-Mining is pleased to introduce the new book, Principles of Database Management – The Practical Guide to Storing, Managing and Analyzing Big and Small Data, by Lemahieu W., vanden Broucke S.,  and Baesens B. (ISBN: 9781107186125). The following is the book interview.




Why did you decide to write your book Principles of Database Management?

This book is the result of having taught an undergraduate database management class and a postgraduate advanced database management class for more than ten years at KU Leuven (Belgium).  Throughout these years, we have found no good textbook which covers the material in a comprehensive way without getting flooded by theoretical detail and losing focus.  Hence, after having teamed up together, we decided to start writing a book ourselves.  This work aims to offer a complete and practical guide covering all the governing principles of database management, including:
  •          End-to-end coverage starting with legacy technologies to emerging trends such as Big Data, NoSQL databases, Analytics, data governance, etc.
  •        A unique perspective on how lessons learnt from past data management could be relevant in today’s technology setting (e.g., navigational access and its perils in Codasyl and XML/OO databases)
  •          A critical reflection and accompanying risk management considerations when implementing the technologies considered, based on our own experiences from participating in data and analytics related projects with industry partners in a variety of sectors, from banking to retail and from government to the cultural sector
  •          Offering a solid balance between theory and practice, including various exercises, industry examples and case studies originating from a diversified and complimentary business practice, scientific research and academic teaching experience


We hear a lot of companies complaining about bad data quality these days.  How can database management contribute to this?
First of all, data quality (DQ) is often defined as ‘fitness for use,’ which implies the relative nature of the concept.  Data of acceptable quality in one decision context may be perceived to be of poor quality in another decision context, even by the same business user.  For instance, the extent to which data is required to be complete for accounting tasks may not be required for analytical sales prediction tasks.  Database management can contribute in various ways to improving data quality. 
A good conceptual data model capturing the business requirements as accurately as possible is the start of everything.  As discussed in the book, both EER and UML can be used for this purpose.  It is important to also list the semantic shortcomings of the conceptual models developed such that they can be followed up during application development. 
A next important activity concerns metadata management.  Just as raw data, also metadata is data that needs to be properly modeled, stored and managed.  Hence, the concepts of data modeling should also be applied to metadata in a transparent way.  In a DBMS approach, metadata is stored in a catalog, sometimes also called data dictionary or data repository, which constitutes the heart of the database system. 
Finally, to manage and safeguard data quality, a data governance culture should be put in place assigning clear roles and responsibilities. The ultimate aim of data governance is to set up a company-wide controlled and supported approach towards data quality, accompanied by data quality management processes. The core idea is to manage data as an asset rather than a liability, and adopt a proactive attitude towards data quality problems. To succeed, it should be a key element of a company’s corporate governance and supported by senior management.

Should all companies start to invest in NoSQL databases?

The explosion of popularity of NoSQL databases should be put in perspective considering their limitations.  Most NoSQL implementations have yet to prove their true worth in the field (most are very young and in development).  Most implementations sacrifice ACID (atomicity, consistency, isolation and durability) concerns in favor of being eventually consistent, and the lack of relational support makes expressing some queries or aggregations particularly difficult, with map-reduce interfaces being offered as a possible, but harder to learn and use, alternative.  Combined with the fact that RDBMSs do provide strong support for transactionality, durability and manageability, quite a few early adopters of NoSQL were confronted with some sour lessons.  It would be an over-simplification to reduce the choice between RDBMSs and NoSQL databases to a choice between consistency and integrity on the one hand, and scalability and flexibility on the other.  The market of NoSQL systems is far too diverse for that.  Still, this tradeoff will often come into play when deciding on taking the NoSQL route.  We see many NoSQL vendors focusing again on robustness and durability.  We also observe traditional RDBMS vendors implementing features that let you build schema-free, scalable data stores inside a traditional RDBMS, capable to store nested, semi-structured documents, as this seems to remain the true selling point of most NoSQL databases, especially those in the document store category.  Expect the future trend to continue towards adoption of such “blended systems”, except for use cases that require specialized, niche database management systems.  In these settings, the NoSQL movement has rightly taught users that the one size fits all mentality of relational systems is no longer applicable and should be replaced by finding the right tool for the job.  For instance, graph databases arise as being “hyper-relational” databases, which makes relations first class citizens next to records themselves rather than doing away with them altogether. 

Many companies are investing in data lakes these days.  What is the difference with a data warehouse?

Much more recent than data warehouses, the data lake concept became known as part of the big data and analytics trend.  Although both data warehouses and data lakes are essentially data repositories, there are some clear differences as listed in the table below.


Data Warehouse
Data lake
Data
Structured
Often unstructured
Processing
Schema-on-write
Schema-on-read
Storage
Expensive
Low cost
Transformation
Before entering the DW
Before analysis
Agility
Low
High
Security
Mature
Maturing
Users
Decision makers
Data Scientists

A key distinguishing property of a data lake is that it stores raw data in its native format, which could be structured, unstructured or semi-structured.  This makes data lakes fit for more exotic and ‘bulk’ data types that we generally do not find in data warehouses, such as social media feeds, clickstreams, server logs, sensor data, etc.  A data lake collects data emanating from operational sources ‘as is’, often without knowing upfront which analyses will be performed on it, or even whether the data will ever be involved in analysis at all.  For this reason, either no or only very limited transformations (formatting, cleansing, …) are performed on the data before it enters the data lake.  Consequently, when the data is tapped from the data lake to be analyzed, quite a bit of processing will typically be required before it is fit for analysis.  The data schema definitions are only determined when the data is read (schema-on-read) instead of when the data is loaded (schema-on-write) as is the case for a data warehouse.  Storage costs for data lakes are also relatively low because most of the implementations are open-source solutions that can be easily installed on low-cost commodity hardware.  Since a data warehouse assumes a predefined structure, it is less agile compared to a data lake which has no structure.  Also, data warehouses have been around for quite some time already, which automatically implies that their security facilities are more mature.  Finally, in terms of users, a data warehouse is targeted towards decision makers at middle and top management level, whereas a data lake requires a data scientist, which is a more specialized profile in terms of data handling and analysis.

How do Object Oriented DBMSs (OODBMSs) compare against Relational Databases (RDBMSs)? Why haven’t they been that successful?

OODBMSs offer several advantages.  First, they store complex objects and relationships in a transparent way.  The identity based approach allows for improved performance when performing complex queries involving multiple interrelated objects, avoiding expensive joins.  By using the same data model as the programming language to develop database applications, the impedance mismatch problem is no longer an issue.  In addition, developers can be more productive, as they are confronted with only a single language and data model.
Still, the success of OODBMSs has been limited to niche applications characterized by complex, nested data structures where an identity-based, instead of a value-based, method of working pays off.  An example is the processing of scientific data sets by CERN in Switzerland, where data access follows predictable patterns.  The widespread use and performance of RDBMSs, however, proved hard to displace: the (ad-hoc) query formulation and optimization procedures of OODBMSs are often inferior to relational databases, which all adopt SQL as their primary database language combined with a powerful query optimizer.  When compared to RDBMSs, OODBMSs are not that developed in terms of robustness, security, scalability and fault-tolerance.  They also don’t provide a transparent implementation of the three-layer database architecture.  More specifically, most OODBMSs provide no support for defining external database models, such as views in the relational model. 

Why are so many companies still struggling with database legacy instead of investing in newer technologies?

Many firms still struggle with legacy databases due to historical implementations and limited IT budgets.  Hence, knowing the basic characteristics thereof is essential to the maintenance of the corresponding database applications and the potential migration to modern DBMSs.  Second, understanding the basics of these legacy models will contribute to better understanding the semantical richness of newer database technologies.  Finally, the concept of procedural DML and navigational access originally introduced by these legacy models has also been adopted by more recent databases such as OODBMSs. 

This book covers a few decades of evolutions in the database world.  Which developments or patterns struck you as the most remarkable in the long run ?
The perseverance of SQL is really striking.  It survived object-oriented databases and XML databases, and after a while, many of its constructs were even incorporated in these paradigms.  More recently, there is considerable irony (and beauty) in the fact that one of the main tendencies of NoSQL development efforts is to mimic the convenient data retrieval and query facilities of, well, SQL-based databases.  Solutions for ‘SQL on top of NoSQL’ are all over the place.  And let’s not forget: SQL still remains a beautiful language.

Is there any additional material provided (e.g., for instructors or students)?
Yes, the book comes with the following additional material:
  • A website with additional information: www.pdbmbook.com
  • Free YouTube lectures for each of the 20 chapters, see https://www.youtube.com/watch?v=o36Z_OqC2ac&list=PLdQddgMBv5zHcEN9RrhADq3CBColhY2hl
  • PowerPoint slides for each of the 20 chapters, see http://www.pdbmbook.com/lecturers
  • A solutions manual with the solutions to all multiple choice and open questions.
  • An online playground with diverse environments, including MySQL for querying; MongoDB; Neo4j Cypher; and a tree structure visualization environment
  • Full-color illustrations throughout the text
  • Extensive coverage of important trending topics, including data warehousing, business intelligence, data integration, data quality, data governance, Big Data, and analytics
  • Hundreds of examples to illustrate and clarify the concepts discussed that can be reproduced on the book’s companion online playground
  • Case studies, review questions, problems, and exercises in every chapter


Wednesday, January 24, 2018

Milestone Project Accomplished- Using Artificial Intelligence to Calculate Characteristic Town Index

West Lake in Hangzhou, China

I published an article today on my Linkedin profile Using Artificial Intelligence to Calculate Characteristic Town Index In this article, I described a milestone project that we recently accomplished and how it made me instantly famous in the field.The follow is the content of the Linkedin article.

Characteristic town development is a hot topic in China. I am extremely excited to announce that using artificial intelligence we have successfully developed Characteristic Town Index for 3,700 towns in China. Based on the index, we are able to identify the most promising towns objectively and efficiently for development. This is the first time that artificial intelligence is used in characteristic town evaluation. It is a cross-team effort under my leadership as the CTO of Hangzhou Jingli Company.

I published an article, Use Big Data and Artificial Intelligence to Rationally Evaluate the Characteristic Town, on influential "Economic Information Daily" (in Chinese) on January 10, 2018 (Please use Google Translate if you are interested). It quickly becomes one of the most cited articles on the Internet in China in this field. On one website alone, the article has been read nearly 52 thousand times. Websites hosting the article consistently ranked number one by Google and Baidu on keywords "characteristic town assessment artificial intelligence" (in Chinese). After reading the article, a government official of a major city in China says, "Dr. Zhou's article should be read carefully and thoroughly. The points raised are thought-provoking". The article's main point is that data-driven artificial intelligence models have advantages over experience-based expert systems. The following are some highlights:
  • Multiple teams have spent months to collect 69 variables that are relevant for 3,700 towns. These variables cover climate, geography, economy, ecology, transportation, Internet infrastructure and so on.
  • Expert systems are subjective, rigid and static. It is hard to adjust expert systems based on the discrepancies between their outputs and the realities. The feedback mechanism is lacking.
  • Data, machine learning models and applications should form closed loops. The data are used to train models. Models' outputs provide decision support for applications. The data are updated and enhanced when models are applied in real world, e.g., finding better target variables. These closed loops allow models to be improved iteratively.

This article has made me instantly famous in this field in China. However, the success doesn't come instantly at all. The core ideas are from my award-winning academic paper published two decades ago "Using genetic learning neural networks for spatial decision making in GIS". I feel fulfilled that I finally implement the ideas for practical applications in the field of GIS.