此数据源类型不支持varsToKeep和varsToDrop的rxRocCurve错误 [英] rxRocCurve errors with varsToKeep and varsToDrop not supported for this data source type

问题描述

大家好，

我正在执行SQL Server R Services数据科学端到端演练（https://msdn.microsoft.com/en-us /library/mt629164.aspx)

I am executing the SQL Server R Services data science end-to-end walkthrough (https://msdn.microsoft.com/en-us/library/mt629164.aspx)

执行以下语句时：

rxRocCurve（actualVarName =" tipped"，predVarNames =" Score" ，data = scoredOutput）

rxRocCurve(actualVarName="tipped", predVarNames="Score", data=scoredOutput)

我遇到以下错误：

'varsToKeep' and 'varsToDrop' not supported for this data source type.
Warning messages:
1: In rocComputeWithTransformObjects(actualVarName = actualVarName,  :
  There are no observations with actual 'TRUE' values. 'sensitivity' cannot be computed.
2: In rocComputeWithTransformObjects(actualVarName = actualVarName,  :
  There are no observations with actual 'FALSE' values. 'specificity' cannot be computed.

似乎rxRocCurve正在使用变换来改变"倾斜"中的值0,1。为TRUE / FALSE。但是，'varsToKeep'和'varsToDrop'似乎不支持RxSqlServerData数据源的数据源类型。

It appears that rxRocCurve is using transform to change the values 0,1 in "tipped" to TRUE / FALSE. However 'varsToKeep' and 'varsToDrop' does not seem to support the data source type of RxSqlServerData data sources.

是否有其他人遇到此问题？这是一个错误吗？

Is anyone else experiencing this issue? Is this a bug?

我能想到的唯一解决方法是将其更改为本地环境。这是预期的吗？

The only workaround I can think of is to change this to local context. Is this expected?

感谢您的回复和帮助！

最诚挚的问候，

Ian

推荐答案

Ian，

我遇到了你所描述的错误，我正在使用SQL R Services 9。 x，你有这种经历9个月后。示例演练尚未更改为解决此问题，也未在rxRocCurve中修复此问题。

I got the same error you described, and I'm using SQL R Services 9.x, 9 months after you had this experience. The example walkthrough has not been changed to address this issue, nor has the problem been fixed in rxRocCurve.

在我的情况下，有什么不同之处在于我已将计算上下文更改为"local"。 （与"RxLocalSeq"相同），但我也得到了这个错误。

What's different, in my case, is that I had changed the compute context to "local" (same as "RxLocalSeq"), yet I also get this error.

然后，我注意到这段代码稍微向下，这解决了问题（当计算上下文是"我将该代码移到rxRocCurve调用之上后显示; local"）

Then, I noticed this bit of code a little further down, which solved the problem (when compute context is "local") after I moved that code above the call to rxRocCurve.

＃首先，你需要使用rxImport将得分的输出数据带到客户端

# First of all you need to bring the scored Output data to the client using rxImport

scoredOutput = rxImport（scoredOutput）

scoredOutput = rxImport(scoredOutput)

这篇关于此数据源类型不支持varsToKeep和varsToDrop的rxRocCurve错误的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！