以向量为参数的 RSQLite 参数化查询

Question

我不熟悉SQL 及其语法，我无法理解如何使用RSQLite 将多个值（例如向量或列表）传递给R 中的参数化查询中的单个参数。

我有一个包含来自 RNASeq 数据的数据的两个表数据库 (myTCGA)。第一个（tcga_P）包含一些基因样本的表达值（FPKM），而（tcgaMeta）包含这些样本的元数据信息。

#tcga_P
      FPKM           Sample                 Tissue GeneName
5550 0.0633 TCGA-AB-2803-03A Acute_Myeloid_Leukemia  PLEKHN1
5551 0.2390 TCGA-AB-2805-03A Acute_Myeloid_Leukemia  PLEKHN1
5552 0.0253 TCGA-AB-2806-03A Acute_Myeloid_Leukemia  PLEKHN1
5553 0.0385 TCGA-AB-2807-03A Acute_Myeloid_Leukemia  PLEKHN1
5554 0.0326 TCGA-AB-2808-03A Acute_Myeloid_Leukemia  PLEKHN1
5555 0.2836 TCGA-AB-2810-03A Acute_Myeloid_Leukemia  PLEKHN1

# tcgaMeta (only few columns)
SampleIndex         SampleID        SubjectID Tumor.Type         Sample.Type
  1           0 TCGA-01-0628-11A TCGA-01-0628         OV Solid Tissue Normal
  2           1 TCGA-01-0630-11A TCGA-01-0630         OV Solid Tissue Normal
  3           2 TCGA-01-0631-11A TCGA-01-0631         OV Solid Tissue Normal

我想从tcga_P 中仅提取属于特定组的样本（例如所有肺样本）的表达值。为此，我编写了一个如下所示的查询

library(DBI)
library(RSQLite)
library(data.table)

myGene <- "PLEKHN1"
myTissue <- "lung"
myCancer <- "Lung Adenocarcinoma"
selectedSamples <- dbGetQuery(myTCGA, 
     "SELECT A.*
     FROM tcga_P A 
     WHERE A.GeneName = $gene AND
           A.Sample in (SELECT B.SampleID FROM tcgaMeta B 
                WHERE B.Tissue = $tissue AND 
                      B.`Disease.TCGA.` = $cancer )
     ",param = list(gene=myGene,tissue=myTissue,cancer=myCancer))
 # from long to wide
 selectedSamplesWide <- dcast(selectedSamples,GeneName~Sample, value.var = "FPKM",fun.aggregate = sum)

仅当我当时提供一个基因时，此查询才能正确返回值，但如果我尝试从多个基因中提取值，它将失败 (myGene <- c("PLEKHN1","PSMD12"))

Error in result_bind(res@ptr, params) : 
Parameter 2 does not have length 2.

我可以循环 (sapply) 遍历向量中的基因，让 sql 一次调用一个基因并将结果绑定在一起，但我想在 sql 调用中做所有事情。

到目前为止，我尝试过

WHERE A.GeneName IN ($gene)
WHERE A.GeneName IN (SELECT C.GeneName FROM $gene C)

我还尝试将myGene 转换为data.frame 并将基因视为一列。说没有任何效果是多余的。

我错过了什么？参数是怎么传给param = list()的？

Answer 1

在 SQL 中，具有相等性的 WHERE 子句 = 在您尝试将表达式计算为两个值时需要一个值。但是，WHERE 和 IN 允许多个值：

WHERE A.GeneName IN ('PLEKHN1', 'PSMD12', ...)

对于无限数量的值，请考虑使用paste + collapse 动态创建准备好的语句，并使用setNames 和as.list 绑定参数值：

myGene <- c("PLEKHN1", "PSMD12")
myTissue <- "lung"
myCancer <- "Lung Adenocarcinoma"

myPlaceHolders <- paste0("$gene", seq_along(myGene))

sql <- paste0("SELECT A.*
               FROM tcga_P A 
               WHERE A.GeneName IN (", paste(myPlaceHolders, collapse=", "), ") 
                 AND A.Sample in (SELECT B.SampleID 
                                  FROM tcgaMeta B 
                                  WHERE B.Tissue = $tissue 
                                    AND B.`Disease.TCGA.` = $cancer)
             ")

myGeneParams <- as.list(setNames(myGene, gsub("\\$", "", myPlaceHolders)))
paramList <- c(myGeneParams, tissue=myTissue, cancer=myCancer)

selectedSamples <- dbGetQuery(myTCGA, sql, param = myParamList)

Rextester demo（对于准备好的语句和参数输出不查询运行）