I am new to SQL
and its syntax, and I cannot understand how to pass multiple values (e.g. vector or list) to a single parameter in a parametrised query in R
using RSQLite
.
I have a two tables database (myTCGA
) with data coming from RNASeq data. The first (tcga_P
) contains expression values (FPKM
) of samples for some genes, while (tcgaMeta
) contains the metadata information of those samples.
#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
I want to extract the expression values from tcga_P
of only the samples that belong to specific groups (e.g. all lung samples). To do so, I wrote a query that looks like this
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.` = $cancer )
",param = list(gene=myGene,tissue=myTissue,cancer=myCancer))
# from long to wide
selectedSamplesWide <- dcast(selectedSamples,GeneName~Sample, value.var = "FPKM",fun.aggregate = sum)
This query correctly returns the values only if I supply one gene at the time, but it fails if I try to extract the values from multiple genes (myGene <- c("PLEKHN1","PSMD12"))
Error in result_bind(res@ptr, params) :
Parameter 2 does not have length 2.
I could loop (sapply
) through the genes in the vector, make the sql call one gene at the time and bind the results together, but I would like to do everything inside the sql call.
So far I tried with
WHERE A.GeneName IN ($gene)
WHERE A.GeneName IN (SELECT C.GeneName FROM $gene C)
I also tried to convert myGene
to a data.frame
and treat the genes as a column. It is superfluous to say that nothing worked.
What am I missing? how are the parameters passed to param = list()
?
In SQL, the WHERE
clause with equality, =
, expects one value where you are attempting to evaluate expression to two values. However, the WHERE
with IN
allows for multiple values:
WHERE A.GeneName IN ('PLEKHN1', 'PSMD12', ...)
For an open-ended number of values, consider dynamically creating the prepared statement with paste
+ collapse
and binding parameter values with setNames
and as.list
:
myGene <- c("PLEKHN1", "PSMD12")
myTissue <- "lung"
myCancer <- "Lung Adenocarcinoma"
myPlaceHolders <- paste0("$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 (for prepared statement and parameter output not query run)