将矩阵字符值转数值并保留索引转为xts对象问题

问题根源 #

你的矩阵元素是带前后空格和千分位逗号的字符串（比如" 112,055.39 "），直接修改storage.mode时，R无法识别这些非数值字符，导致全部转为NA。

分步解决方案 #

1. 清理字符串并转换为数值矩阵 #

先处理每个字符串的冗余字符，再转数值，同时保留原行列索引：

# 加载依赖包
library(dplyr)
library(tibble)
library(xts)

# 清理并转换：去除空格、替换逗号、转数值
master_zillow3_clean <- master_zillow3 %>%
  as.data.frame() %>%
  rownames_to_column(var = "date") %>%  # 保留原行名（日期）
  mutate(across(-date, ~ as.numeric(gsub(",", "", trimws(.x))))) %>%  # 清理字符串并转数值
  column_to_rownames(var = "date")  # 恢复行名

# 转为数值矩阵
master_zillow3_num <- as.matrix(master_zillow3_clean)

2. 转换为xts对象 #

xts要求行索引为时间类型，需先将原日期字符串转为标准Date格式：

# 把行名（日期字符串）转为Date类型
date_index <- lubridate::mdy(rownames(master_zillow3_num))

# 构造xts对象
master_zillow_xts <- xts(x = master_zillow3_num, order.by = date_index)

完整测试代码（用你的示例数据） #

# 示例数据
sample_data <- structure(c(" 112,055.39 ", " 112,074.39 ", " 111,940.63 ", " 111,935.54 ", 
" 111,989.44 ", " 221,254.91 ", " 221,443.20 ", " 221,951.01 ", 
" 222,971.84 ", " 222,991.41 ", " 92,921.09 ", " 92,903.53 ", 
" 92,922.18 ", " 92,896.29 ", " 92,921.62 ", " 272,971.91 ", 
" 275,503.86 ", " 277,680.63 ", " 281,073.80 ", " 283,666.72 ", 
" 247,208.49 ", " 248,297.55 ", " 249,632.64 ", " 251,537.16 ", 
" 252,856.62 "), dim = c(5L, 5L), dimnames = list(c("1-31-2000", 
"2-29-2000", "3-31-2000", "4-30-2000", "5-31-2000"), c("77449", 
"77494", "79936", "11368", "11385")))

# 执行清理和转换
sample_clean <- sample_data %>%
  as.data.frame() %>%
  rownames_to_column("date") %>%
  mutate(across(-date, ~ as.numeric(gsub(",", "", trimws(.x))))) %>%
  column_to_rownames("date") %>%
  as.matrix()

# 转xts
sample_xts <- xts(sample_clean, order.by = lubridate::mdy(rownames(sample_clean)))

# 查看结果
head(sample_xts)

关键细节 #

• trimws()：去除字符串前后的空格
• gsub(",", "", .x)：删除千分位逗号，让R能识别为数值
• lubridate::mdy()：将"月-日-年"格式的字符串转为标准Date类型，确保xts的时间索引有效

内容的提问来源于stack exchange，提问作者js80