英文:
Issue with ingest values, 2x more
问题
当我将值添加到特征集时，流水线被调用了两次更多（我使用的是 MLRun 版本 1.2.1）。看起来像是个问题，你知道为什么吗？
我使用了这段代码：
import mlrun
import mlrun.feature_store as fstore
# mlrun: start-code
import math
def calc(x):
x['fn2'] = math.sin(x['fn2']) * 100.0
print('calc')
return x
# mlrun: end-code
mlrun.set_env_from_file("mlrun-nonprod.env")
project = mlrun.get_or_create_project(project_name, context='./', user_project=False)
feature_derived = fstore.get_feature_set(f"{project_name}/{feature_derivedName}")
# ...
# dataFrm 只有两个值
feature_derived.graph.to(name="calc", handler='calc')
fstore.ingest(feature_derived, dataFrm)
我得到了以下输出（calc 方法被调用了四次），针对只有两个值的 dataFrm：
> calc
> calc
> calc
> calc
英文:
When I ingested values to the feature set, the pipeline was called 2x more (I used MLRun version 1.2.1). It seems as the issue, do you know why?
I used this code:
import mlrun
import mlrun.feature_store as fstore
# mlrun: start-code
import math
def calc(x):
x['fn2']=math.sin(x['fn2'])*100.0
print('calc')
return x
# mlrun: end-code
mlrun.set_env_from_file("mlrun-nonprod.env")
project = mlrun.get_or_create_project(project_name, context='./', user_project=False)
feature_derived = fstore.get_feature_set(f"{project_name}/{feature_derivedName}")
...
# dataFrm has only two values
feature_derived.graph.to(name="calc", handler='calc')
fstore.ingest(feature_derived, dataFrm)
I got this output (method calc was called four times) for dataFrm with two values:
> calc
> calc
> calc
> calc
答案1
得分: 0
解决方法很简单，只需在数据摄取方法中设置infer_options=0来关闭预览模式。请查看代码的一部分：
...
feature_derived.graph.to(name="calc", handler='calc')
fstore.ingest(feature_derived, dataFrm, infer_options=0)
...
输出仅包含两个值（如所请求）：
> calc
> calc
英文:
The solution is easy, it is enough to switch-off preview mode based on setting infer_options=0 in ingest method. See part of the code
...
feature_derived.graph.to(name="calc", handler='calc')
fstore.ingest(feature_derived, dataFrm, infer_options=0)
...
The output has only two values (as requested):
> calc
> calc
通过集体智慧和协作来改善编程学习和解决问题的方式。致力于成为全球开发者共同参与的知识库，让每个人都能够通过互相帮助和分享经验来进步。
评论