FAQ

Q: What are the differences between the files in the formats such as CSV, SQL, DDL, CUT, POS, and MIX mentioned in this document?

A:

Q: What is the relationship between the format of a data file and the extension of the file?

A: All data files are named in the table.group.sequence.suffix format.

OBLOADER supports six data file formats: CSV, SQL, DDL, CUT, POS, and MIX. The default extension of a file in the CSV format is .csv. The default extension of a file in the SQL format is .sql. The default extension of a file in the DDL format is -schema.sql. You can use any extension for all data formats, except the DDL format.

For example, you can use test.dat as the name of a file in the CSV format, but the content of the file must be in the standard CSV format. When you import a file, you must specify its extension by adding the --file-suffix '.dat' option to the command. Otherwise, an error will be reported, indicating that the data file is not found.

Q: After data in a large table is exported by partition into several files with indefinite extensions, how can I import them by using OBLOADER?

A: You can specify a regular expression for the file extensions in the command. For example, if the names of the data files that you want to import are test.tbl.0, test.tbl.1, and test.tbl., you can specify the following options in the command: --table test --file-suffix '^.{1}\\w+.{1}\\d+'. Make sure that the specified regular expression is correct. Otherwise, an error will be reported, indicating that the data file is not found.

Q: What do I need to pay attention to when I use OBLOADER to import data that was exported by using a third-party tool such as Mydumper or SQL Developer?

A:

Q: What do I need to pay attention to when I specify the --table option?

A: The--table option specifies table names. Separate multiple names with commas (,). The table name and the file name must be capitalized in the same way. For example, in MySQL mode, table names are in lowercase by default. In this case, the table name specified by the--table option must be in lowercase, and the file name specified by the-f option must also be in lowercase. In Oracle mode, table names are in uppercase by default. In this case, the table name specified by the --table option must be in uppercase, and the file name specified by the -f option must also be in uppercase.

Q: What is the difference between using OBLOADER in a public cloud and a private cloud?

A: In the public cloud (limited) mode, the proxyro permissions are limited, but they are not limited in the private cloud mode. The connection methods in these two cloud environments are also different:

In a public cloud, OBLOADER is connected by using the --public-cloud option.

In a private cloud, OBLOADER is connected by using the --sys-password option or a key file. Note

The limited mode (public cloud connection mode) does not support all of the features of OBLOADER. The performance and stability are also not as good as in regular mode. In addition, only OceanBase Database V2.2.30 and later support throttling on the server. Therefore, when you use the limited mode, run the following command to modify the throttling settings on the server:

alter system set freeze_trigger_percentage=50;
alter system set minor_merge_concurrence=64;
alter system set writing_throttling_trigger_percentage=80 tenant='xxx';

Q: I got garbled characters after truncating a long string of Chinese characters. How can I solve this issue?

A:

Correctly use GBK for character encoding. If the application can correctly distinguish carriage returns from line feeds, 90% of the problems would be solved. Note

Dirty data may not be simulated for customers.

Q: OBLOADER imports data to the incorrect table. What can I do?

A: * OBLOADER & OBDUMPER V2.1.14 and later are not allowed to export data to a non-empty directory.

• If the default data export directory is empty, the --skip-check-dir parameter is added to skip the empty-directory check.

  Bug example: Note

  Prerequisite: The command-line option is specified as -f 'directory'. If the command-line option is specified as -f 'a specific file', erroneous writes will not occur.