为何在数据处理管道末尾直接调用gt()会将DataFrame转为列表？

问题现象 #

使用gt包生成表格时出现如下差异：

• 直接在数据处理管道末尾调用gt()并赋值给变量，表格不会在RStudio查看器中显示，且变量存储的gt对象在控制台打印时呈现为列表形式
• 先将处理后的数据存入变量，再单独执行变量 %>% gt()，表格能正常在查看器显示

无法正常显示的代码 #

library(nflfastR)
library(gt)
library(tidyverse)

pbp_rp <- load_pbp(2023) %>%
  drop_na(yards_gained) %>% 
  filter(rush == 1 | pass == 1, !is.na(yards_gained)) %>%
  select(posteam, yards_gained, week) %>%
  group_by(posteam) %>% 
  summarise(att = sum(!is.na(yards_gained)),
            lost_att = sum(yards_gained <0),
            zero_att = sum(yards_gained == "0"),
            positive_att = sum(yards_gained >0),
            lost_yards = sum(yards_gained [yards_gained<0]),
            yards_0_10 = sum(yards_gained [yards_gained>=0 & yards_gained<10]),
            yards_10_20 = sum(yards_gained [yards_gained>=10 & yards_gained<20]),
            yards_20 = sum(yards_gained [yards_gained>=20]),
            positiv_yards = sum(yards_gained [yards_gained>0]),
            yards_total = sum(yards_gained),
            lower = min(yards_gained),
            avg = mean(yards_gained),
            upper =max(yards_gained)) %>% 
  ungroup() %>% 
  gt()

可以正常显示的代码 #

library(nflfastR)
library(gt)
library(tidyverse)

pbp_rp <- load_pbp(2023) %>%
  drop_na(yards_gained) %>% 
  filter(rush == 1 | pass == 1, !is.na(yards_gained)) %>%
  select(posteam, yards_gained, week) %>%
  group_by(posteam) %>% 
  summarise(att = sum(!is.na(yards_gained)),
            lost_att = sum(yards_gained <0),
            zero_att = sum(yards_gained == "0"),
            positive_att = sum(yards_gained >0),
            lost_yards = sum(yards_gained [yards_gained<0]),
            yards_0_10 = sum(yards_gained [yards_gained>=0 & yards_gained<10]),
            yards_10_20 = sum(yards_gained [yards_gained>=10 & yards_gained<20]),
            yards_20 = sum(yards_gained [yards_gained>=20]),
            positiv_yards = sum(yards_gained [yards_gained>0]),
            yards_total = sum(yards_gained),
            lower = min(yards_gained),
            avg = mean(yards_gained),
            upper =max(yards_gained)) %>% 
  ungroup()  
pbp_rp %>% gt()

原因解释 #

核心在于R的赋值操作与对象渲染逻辑差异：

• 当把gt()结果直接赋值给变量时，R仅完成对象存储，不会触发gt包的表格渲染机制，因此查看器无输出；你看到的"列表"是R对gt对象的默认打印格式，并非真的将DataFrame转为列表。
• 单独执行pbp_rp %>% gt()时，表达式未被赋值，R会在控制台返回gt对象，此时gt包的S3方法自动触发渲染，将表格输出到查看器。

解决方法 #

若要在管道末尾赋值的同时显示表格，可在最后追加%>% print()强制触发渲染：

pbp_rp <- load_pbp(2023) %>%
  # 保留原有数据处理步骤
  ungroup() %>% 
  gt() %>% 
  print()

内容的提问来源于stack exchange，提问作者Jackie Nielsen