ChatGPT has been very hot recently, it has passed a lot of exams, so can it replace the human Oracle DBA, test it to see if it can pass the Oracle 19c OCP exam put, because Mr. Yao Yuan is the Oracle OCP and MySQL OCP instructor, so it is very curious about this thing. Let s take the real Oracle 19c OCP exam (1z0-082) questions to ChatGPT for a test. Yao Yuan OracleACE: Try MySQL 8.0 Operation and Maintenance and Optimization (Tsinghua University Press) 0 Agree - 0 Comment Article Test Process The 1st question is about the use of space in the Oracle database, ChatGPT made a mistake, Yao Yuan teacher can not help but feel happy in his heart, it seems that ChatGPT is not very good, and we may not be out of work, let's take a look at the first Question Title:Which two statements are true about space-saving features in an Oracle Database?A. An index created with the UNUSABLE attribute has no segment. Private Temporary Tables(PTTs) store metadata in memory only.C.Private Temporary Tables(PTTs) when used, are always dropped at the next commit statement.D.An index that is altered to be UNUSABLE will retain its segmentE. A table that is truncated will always have all of its extents removedChatGPT The answers to the choices are B and C. B is right and C is wrong: Private Temporary Tables (PTTs) when used, are always dropped at the next commit statement. There are actually two types of PTTs: Transaction-specific ( There are actually two types of PTTs: transaction-specific (ON COMMIT DROP DEFINITION) and session-specific (ON COMMIT PRESERVE DEFINITION). Both of them will delete the data in the table when COMMIT, but the second one will not delete the table definition, so C is wrong.ChatGPT left out a correct answer A, so Mr. Yao Yuan asked ChatGPT: Why didn't you choose A, could you please do an experiment and try to create an UNUSABLE index? ChatGPT replied: as an AI language model, I don't have a direct physical database to create an index... Now I know why ChatGPT is doing it wrong, because ChatGPT doesn't have an experimental environment. I experimented myself: SQL> create index in_1 on table_1(no) unusable; Index created. SQL> select segment_name,blocks from user_segments where segment_name in (' SQL> create index in_1 on table_1(no) unusable; Index created. SQL> select segment_name,blocks from user_segments where segment_name in ('IN_1'); no rows selected SQL> alter index in_1 rebuild; Index altered. 1'); SEGMENT_NAME BLOCKS -------------------- ---------- IN_1 8 Found out that creating an index as unusable does not create segments, only rebuilding the index does. The 2nd question was about views, ChatGPT was clear about the WITH CHECK restriction and got it right. Question 3 is about PMON process assignments, ChatGPT also knows that the function of PMON registering and listening has been given to the LRRG process to take care of since 12c, and got it right as well. Question 4 was the INTERSECT operation for collections, which ChatGPT also got right. Question 5 was about the principle of least privilege for users, which ChatGPT also got right. Question 6 is about rollback segments, and ChatGPT got it wrong. ChatGPT thinks that Undo segments can be stored in the SYSTEM tablespace. ChatGPT gives the following reason: Undo segments are stored in the undo tablespace, not in the SYSTEM tablespace. In fact, in the automatic rollback segment management, if there is no undo tablespace, the rollback segments can be placed in the SYSTEM tablespace, in the Oracle's official document there is the following content: When the database instance starts When the database instance starts , the database automatically selects the first available undo tablespace. If no undo tablespace is available, then the instance starts without an undo If no undo tablespace is available, then the instance starts without an undo tablespace and stores undo records in the SYSTEM tablespace. This is not recommended, and an alert message is written to the alert log file to warn that the This is not recommended, and an alert message is written to the alert log file to warn that the system is running without an undo tablespace. Question 7 was a date calculation, which I also got right. ChatGPT chose the following option: Directory Naming requires setting the TNS_ADMIN environment variable on the client side. In fact, the TNS_ADMIN environment variable is not required to be set. In fact, the TNS_ADMIN environment variable is not required to be set, only ORACLE_HOME is set. ChatGPT chose the wrong answer for question 9: Any user can create a PUBLIC synonym. In fact, even creating a PUBLIC synonym requires the CREATE PUBLIC SYNONYM system privilege. ChatGPT also ignored the correct answer: A synonym can have a synonym. the 10th question is about the compressed format of direct path import, this knowledge is written clearly in the official Oracle documentation, ChatGPT got it right. the 11th question is about the delayed segment creation feature, ChatGPT also got it right. summarize and Analysis of the test carried out here, ChatGPT did a total of 11 questions, wrong 4 questions, the correct rate of 63.6%, and Oracle 19C OCP's 1Z0-082 pass line is 60%, ChatGPT waded through!