apache/doris

AI Summary: When a `CREATE ROUTINE LOAD` job includes a `CAST` function in its column definitions (e.g., `CAST(tmp/60 AS int)`), the `SHOW ROUTINE LOAD` command incorrectly displays the cast target type as `INVALID_TYPE` instead of the actual type (e.g., `int`). The expected behavior is for the `SHOW` command to accurately reflect the `CAST` function's target type.

AI Summary: This GitHub issue suggests an improvement to the Apache Doris documentation. It highlights that the `be_config.md` page, which details Backend (BE) configurations, is missing instructions for most of its entries. The user requests that explanations be added for these configurations to help users understand their meaning and proper usage.

AI Summary: The backend (BE) component of the system occasionally crashes with a `SIGABRT` signal. The crash log indicates an assertion failure within `pthread_mutex_lock`, specifically `mutex->__data.__owner == 0' failed. The stack trace points to `doris::OlapScanNode::scanner_thread()` as the function where the problematic mutex lock operation occurs, suggesting a potential race condition or incorrect mutex usage.

AI Summary: The user reports a "cast error" when querying a partitioned table. The provided SQL demonstrates a type mismatch in the `WHERE` clause, where the integer output of the `DAYOFWEEK()` function is being compared against a string literal ('2') using the `IN` operator, which is the likely cause of the error.

AI Summary: A user is encountering 'TOO_MANY_TASKS' errors when using ROUTINE LOAD to ingest data from Kafka into Doris, causing all load tasks to fail. Despite increasing `max_routine_load_task_num_per_be` and `max_routine_load_job_num` configurations and restarting the frontend node, the issue persists. They are seeking further solutions to resolve the routine load failures.

AI Summary: The user requests an enhancement to the Doris documentation on "running profiles." While the current documentation explains what profiles are, the user wants practical examples demonstrating how to adjust configurations using profiles to achieve high performance. This involves adding concrete configuration adjustment scenarios to the existing documentation.

AI Summary: The user reports that the `HELP SHOW LOAD` command's documentation is incomplete. It currently fails to inform users that `SHOW LOAD` does not display stream load tasks. The expected behavior is for the help text to be updated to include this crucial clarification.

AI Summary: The user is inquiring about the current stability of the 'Multi Tenant' feature. They note that it has been marked as experimental in the documentation for an extended period and are seeking an update on its status.

AI Summary: A Doris user is experiencing a Backend (BE) crash in a CentOS 7.5 environment when performing a second `INSERT` operation into a table with a `bitmap_union` column, followed by a `SELECT` query. The error message indicates "Lost connection to MySQL server during query" and logs show "failed to get largest_segment_group". The same operations work without issues in a CentOS 7.8 environment.

AI Summary: The user is inquiring whether Apache Doris supports a utility similar to `mysqldump` for exporting the schema, specifically the table structures, of an entire database. They are seeking information on how to perform a database schema export in Doris.

AI Summary: The user reported that the `REPLACE()` string function, commonly found in MySQL, is not available in Doris. Attempting to use it results in a syntax error, indicating that `REPLACE` is recognized as a keyword but not as a callable function.

AI Summary: The user is asking if configuring multiple disks into a RAID array can improve disk I/O performance. They are comparing this to using multiple independent disks for a single backend or block element and want to understand RAID's impact on I/O.

AI Summary: The user is testing Doris and is looking for functionality to update table data, similar to an `UPDATE` statement in traditional databases. They have observed DML/DDL operations but cannot find a direct way to update existing rows. They are requesting documentation or an explanation of how to perform such updates, noting the `broker load`'s `merge` option as a potential, but unclear, method.

AI Summary: This feature request asks Apache Doris to support the SQL `UPDATE` statement, which is currently not supported and throws a syntax error. The user suggests looking at Apache Impala as an example for implementation, noting that `UPDATE` is a standard feature in most SQL databases.

AI Summary: The issue reports a bug where connecting to Doris via a MySQL client or JDBC and executing `select version()` returns an incorrect version string of "5.1". The expected behavior is for it to return "0.12", which is likely the actual Doris version. This indicates a misrepresentation of the product version.

AI Summary: The user is encountering an issue during stream loading where the default line break character (\n) is incorrectly splitting records because their data also contains \n within fields. This leads to records being parsed with an incorrect number of columns, resulting in a "columns number doesn't match error." The user is asking how to configure a different line break delimiter for stream loading.

AI Summary: The user is reporting a build failure for the `fe-core` module in version 0.13, specifically due to a `Failed to read artifact descriptor` error for the `net.sourceforge.czt.dev:cup-maven-plugin:jar:1.6-cdh` dependency. This indicates that Maven cannot locate or download the necessary plugin artifact.

AI Summary: The issue proposes adding `clang-format` and `clang-tidy` support to the Doris project to improve static code formatting and checking, which is currently lacking. The issue creator intends to provide a patch to implement this feature.

AI Summary: The user is inquiring about Doris's support for the ARM platform. They want to know if it's supported, how to build it for ARM, if it functions correctly, and what modifications would be necessary if it currently lacks support.

AI Summary: This enhancement requests the implementation of the `KILL ANALYZE` command within the `nereids` (new Doris FE planner). Currently, this command, which allows terminating analysis jobs by ID, is only supported by the legacy planner. The task involves adding support for `KILL ANALYZE jobId=INTEGER_VALUE` in `nereids`.

AI Summary: The user is reporting a compilation failure of Apache Doris on CentOS 7, following the official source installation guide using the `ldb_toolchain`. The build process specifically failed when compiling the `minizip` dependency, indicated by a failing `libtool` command. The exact error message is truncated, but points to an issue within the `minizip` build step.

AI Summary: The user reports a syntax error in the `help broker load` command's example for loading data from an HDFS HA cluster in Doris version 0.15.0. The example provided for `LOAD WITH HDFS` is missing a comma after the `"hdfs_user"="user"` configuration, leading to a grammar problem.

AI Summary: The user is reporting a Doris BE (Backend) cluster crash and is unable to analyze the generated core dump effectively using GDB. GDB reports "Cannot access memory" errors and fails to read the object file image, ultimately terminating with a Segmentation fault and an incomplete backtrace pointing to `stl_bvector.h`. The user is asking for assistance in diagnosing the issue.

AI Summary: This enhancement issue proposes replacing existing `Substitute` `stringstream` usages with `fmt/format` in the Backend (BE) component of the project. The goal is to modernize and potentially improve string formatting practices across the codebase.

AI Summary: The user encountered a build error with Doris Manager 1.0.0_rc03 when using Spring Boot 2.5.2. They resolved the build issue by upgrading to Spring Boot 2.6.3, but now face a runtime error due to a circular dependency among `paloUserSpaceController`, `paloUserSpaceService`, and `dorisManagerUserSpaceComponent` in the application context.

AI Summary: The user is experiencing an issue with the `flink-doris-connector` where data loaded into Doris via stream load is reported as successful, but the data is not readable from the Doris table. This suggests a potential problem with data visibility, consistency, or the connector's interaction with Doris's data storage.

AI Summary: A user reports that a mixed deployment of Doris BEs (one ARM, five x86) in version 1.1.3 leads to instability. Under high-concurrency queries (50 concurrent, returning millions of rows), two specific x86 BEs consistently crash, with GDB pointing to a memory allocation error. The problem is resolved by removing the ARM BE, suggesting a compatibility issue in mixed-architecture environments.

AI Summary: When importing data that includes scientific notation (e.g., `1.23E+05`) into a `DECIMAL` column in Apache Doris, the system incorrectly reports an "Unknown column 'E' in 'table list'" error. The expected behavior is a more descriptive "Type conversion exception," indicating a fundamental parsing issue where the 'E' in scientific notation is misinterpreted, potentially as a column name.

AI Summary: This issue reports a BRPC communication problem on port 8060 in Doris version 1.1.3, leading to RPC call timeouts (60000ms) when attempting to open a tablet writer. The problem prevents the service from running normally. No reproduction steps are provided, but the user is willing to submit a PR.

AI Summary: This GitHub issue reports a bug where DataX fails to import `nvarchar` data from SQL Server into Doris (version 1.1.3), resulting in a "Request failed with code: 500" error. The problem specifically occurs when DataX attempts to read and transfer `nvarchar` data from a SQL Server source to a Doris sink.

AI Summary: After upgrading Doris from version 1.13 to 1.2rc1, users are unable to perform `UPDATE` operations on tables that existed prior to the upgrade. However, if the exact same table creation statement is used to create a *new* table in 1.2rc1, `UPDATE` operations work correctly on these newly created tables. This indicates a potential compatibility or migration issue affecting existing table metadata or internal structure during the upgrade process.

AI Summary: The user reports a significant performance regression (10x slower) when writing real-time data from Flink (consuming Kafka, synced via OGG) to Doris 1.1.3, compared to Doris 1.1.2 or Palo 1.0.5. The Doris BE logs show recurring errors like "not found entity", "merge meet errorInvalid argument", and "apply filter meet error: Invalid argument", indicating issues potentially related to runtime filters or internal service communication during data ingestion.