googleapis / googleapis/google-cloud-java
[java-bigquery] BigQuery: data is sometimes not inserted in tables
- 主要语言
- Java
- 星标
- 2.1k
- 派生
- 1.2k
- 平均合并
- 1 天 23 小时
- 30 天内合并 PR
- 157
描述
Hi,
I'm having random data missing/not showing up in BigQuery issues. I would seemingly write data without issues, but SELECT * would return nothing.
1. Have tried latest version but still same issue.
```
com.google.cloud
google-cloud-bigquery
2.28.0
```
2. Linux and Windows
3. JDK 8.202
#### Steps to reproduce
Issue is random, can't reproduce reliably. The writes are batched to 1-2k rows and 99% they work.
The control panel shows no errors. None of the quotas show up as exceeded. I get no exceptions and response.getInsertErrors() is always null. The issue seems to happen when doing intensive writes, as less batches per minute of the same size work fine.
Any ideas what's wrong? thank you.
#### Code example
```java
GoogleCredential credential = new GoogleCredential.Builder().setTransport(new NetHttpTransport())
.setJsonFactory(new JacksonFactory())
.setClientSecrets("clientId", "clientSecret")
.build()
.setFromTokenResponse(new TokenResponse().setRefreshToken("refreshToken");
Bigquery bigquery = new Bigquery.Builder(new NetHttpTransport(), new JacksonFactory(), credential)
.setApplicationName("name").build();
List rowList = new ArrayList();
TableRow rowWrapper = new TableRow();
rowWrapper.set("date", item.GetDate());
rowWrapper.set("value", item.GetValue());
TableDataInsertAllRequest.Rows insertRowRequest = new TableDataInsertAllRequest.Rows();
rowList.add(insertRowRequest);
TableDataInsertAllRequest content = new TableDataInsertAllRequest().setRows(rowList);
InsertAll op = bigquery.tabledata().insertAll("project", "dataset", "table", content);
TableDataInsertAllResponse response = op.execute();
if (response.getInsertErrors() == null)
{
; // all good
}
// example
贡献指南
调研方向
The report names Bigquery.tabledata().insertAll and TableDataInsertAllResponse.getInsertErrors(), but no repository file or test. Start by building a minimal high-throughput reproduction against BigQuery and inspecting the insert response behavior; done means the failure is isolated with evidence or the issue is documented as unreproducible.
由索引模型根据 Issue 内容生成。
评估
- 技术栈
- google-cloud, java
- 领域
- cloud, databases
- Issue 类型
- 缺陷
- 难度
- 4/5
- 预计耗时
- 3-5 天
- 活跃度
- 停滞
- 描述清晰度
- 需要澄清
- 新手友好度
- 25/100