与sparklyr一起使用时sample_n真的是随机样本吗？

Question

我在spark数据框中有5亿行。我对使用sample_n中的dplyr很感兴趣，因为它可以让我明确指定所需的样本量。如果要使用sparklyr::sdf_sample()，则必须首先计算sdf_nrow()，然后创建数据sample_size / nrow的指定分数，然后将该分数传递给sdf_sample。没什么大不了，但是sdf_nrow()可能需要一段时间才能完成。

因此，直接使用dplyr::sample_n()是理想的。但是，经过一些测试，看起来sample_n()并不是随机的。实际上，结果与head()相同！如果该函数只是返回前n行，而不是随机采样行，那将是一个主要问题。

其他人可以确认吗？ sdf_sample()是我最好的选择吗？

# install.packages("gapminder")

library(gapminder)
library(sparklyr)
library(purrr)

sc <- spark_connect(master = "yarn-client")

spark_data <- sdf_import(gapminder, sc, "gapminder")


> # Appears to be random
> spark_data %>% sdf_sample(fraction = 0.20, replace = FALSE) %>% summarise(sample_mean = mean(lifeExp))
# Source:   lazy query [?? x 1]
# Database: spark_connection
  sample_mean
        <dbl>
1    58.83397


> spark_data %>% sdf_sample(fraction = 0.20, replace = FALSE) %>% summarise(sample_mean = mean(lifeExp))
# Source:   lazy query [?? x 1]
# Database: spark_connection
  sample_mean
        <dbl>
1    60.31693


> spark_data %>% sdf_sample(fraction = 0.20, replace = FALSE) %>% summarise(sample_mean = mean(lifeExp))
# Source:   lazy query [?? x 1]
# Database: spark_connection
  sample_mean
        <dbl>
1    59.38692
> 
> 
> # Appears to be random
> spark_data %>% sample_frac(0.20) %>% summarise(sample_mean = mean(lifeExp))
# Source:   lazy query [?? x 1]
# Database: spark_connection
  sample_mean
        <dbl>
1    60.48903


> spark_data %>% sample_frac(0.20) %>% summarise(sample_mean = mean(lifeExp))
# Source:   lazy query [?? x 1]
# Database: spark_connection
  sample_mean
        <dbl>
1    59.44187


> spark_data %>% sample_frac(0.20) %>% summarise(sample_mean = mean(lifeExp))
# Source:   lazy query [?? x 1]
# Database: spark_connection
  sample_mean
        <dbl>
1    59.27986
> 
> 
> # Does not appear to be random
> spark_data %>% sample_n(300) %>% summarise(sample_mean = mean(lifeExp))
# Source:   lazy query [?? x 1]
# Database: spark_connection
  sample_mean
        <dbl>
1    57.78434


> spark_data %>% sample_n(300) %>% summarise(sample_mean = mean(lifeExp))
# Source:   lazy query [?? x 1]
# Database: spark_connection
  sample_mean
        <dbl>
1    57.78434


> spark_data %>% sample_n(300) %>% summarise(sample_mean = mean(lifeExp))
# Source:   lazy query [?? x 1]
# Database: spark_connection
  sample_mean
        <dbl>
1    57.78434
> 
> 
> 
> # === Test sample_n() ===
> sample_mean <- list()
> 
> for(i in 1:20){
+   
+   sample_mean[i] <- spark_data %>% sample_n(300) %>% summarise(sample_mean = mean(lifeExp)) %>% collect() %>% pull()
+   
+ }
> 
> 
> sample_mean %>% flatten_dbl() %>% mean()
[1] 57.78434
> sample_mean %>% flatten_dbl() %>% sd()
[1] 0
> 
> 
> # === Test head() ===
> spark_data %>% 
+   head(300) %>% 
+   pull(lifeExp) %>% 
+   mean()
[1] 57.78434

Answer 1

不是。如果检查执行计划（optimizedPlan功能已定义为here），您会发现它只是一个限制：

spark_data %>% sample_n(300) %>% optimizedPlan()

<jobj[168]>
  org.apache.spark.sql.catalyst.plans.logical.GlobalLimit
  GlobalLimit 300
+- LocalLimit 300
   +- InMemoryRelation [country#151, continent#152, year#153, lifeExp#154, pop#155, gdpPercap#156], true, 10000, StorageLevel(disk, memory, deserialized, 1 replicas), `gapminder`
         +- Scan ExistingRDD[country#151,continent#152,year#153,lifeExp#154,pop#155,gdpPercap#156] 

show_query进一步证实了这一点：

spark_data %>% sample_n(300) %>% show_query()

<SQL>
SELECT *
FROM (SELECT *
FROM `gapminder` TABLESAMPLE (300 rows) ) `hntcybtgns`

和可视化的执行计划：

最后，如果您选中Spark source，您会发现这种情况是通过简单的LIMIT实现的：

case ctx: SampleByRowsContext =>
  Limit(expression(ctx.expression), query)

我相信这种语义是从Hive where equivalent query takes n first rows from each input split继承的。

实际上，获取精确大小的样本非常昂贵，除非绝对必要，否则应避免使用（与大LIMITS相同）。