R devel - Sep 2017 - readLines() segfaults on large file & question on how to work around

Hi:

I have a 2.1GB JSON file. Typically I use readLines() and
jsonlite:fromJSON() to extract data from a JSON file.

When I try and read in this file using readLines() R segfaults.

I believe the two salient issues with this file are
1). Its size
2). It is a single line (no line breaks)

I can reproduce this issue as follows
#Generate a big file with no line breaks
# In R
> writeLines(paste0(c(letters, 0:9), collapse=""),
"alpha.txt", sep="")
# in unix shell
cp alpha.txt file.txt
for i in {1..26}; do cat file.txt file.txt > file2.txt && mv -f
file2.txt
file.txt; done

This generates a 2.3GB file with no line breaks

in R:
> moo <- readLines("file.txt")
 *** caught segfault ***
address 0x7cffffff, cause 'memory not mapped'

Traceback:
 1: readLines("file.txt")

Possible actions:
1: abort (with core dump, if enabled)
2: normal R exit
3: exit R without saving workspace
4: exit R saving workspace
Selection: 3

I conclude:
 I am potentially running up against a limit in R, which should give a
reasonable error, but currently just segfaults.

My question:
Most of the content of the JSON is an approximately 100K x 6K JSON
equivalent of a dataframe, and I know R can handle much bigger than this
size. I am expecting these JSON files to get even larger. My R code lives
in a bigger system, and the JSON comes in via stdin, so I have absolutely
no control over the data format. I can imagine trying to incrementally
parse the JSON so I don't bump up against the limit, but I am eager for
suggestions of simpler solutions.

Also, I apologize for the timing of this bug report, as I know folks are
working to get out the next release of R, but like so many things I have no
control over when bugs leap up.

Thanks.

Jen
> sessionInfo()
R version 3.4.1 (2017-06-30)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 14.04.5 LTS

Matrix products: default
BLAS: R-3.4.1/lib/libRblas.so
LAPACK:R-3.4.1/lib/libRlapack.so

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C
 [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8
 [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8
 [7] LC_PAPER=en_US.UTF-8       LC_NAME=C
 [9] LC_ADDRESS=C               LC_TELEPHONE=C
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base

loaded via a namespace (and not attached):
[1] compiler_3.4.1

	[[alternative HTML version deleted]]

As s work-around I  suggest readr::read_file.

--Ista


On Sep 2, 2017 2:58 PM, "Jennifer Lyon" <jennifer.s.lyon at
gmail.com> wrote:
> Hi:
>
> I have a 2.1GB JSON file. Typically I use readLines() and
> jsonlite:fromJSON() to extract data from a JSON file.
>
> When I try and read in this file using readLines() R segfaults.
>
> I believe the two salient issues with this file are
> 1). Its size
> 2). It is a single line (no line breaks)
>
> I can reproduce this issue as follows
> #Generate a big file with no line breaks
> # In R
> > writeLines(paste0(c(letters, 0:9), collapse=""),
"alpha.txt", sep="")
>
> # in unix shell
> cp alpha.txt file.txt
> for i in {1..26}; do cat file.txt file.txt > file2.txt && mv -f
file2.txt
> file.txt; done
>
> This generates a 2.3GB file with no line breaks
>
> in R:
> > moo <- readLines("file.txt")
>
>  *** caught segfault ***
> address 0x7cffffff, cause 'memory not mapped'
>
> Traceback:
>  1: readLines("file.txt")
>
> Possible actions:
> 1: abort (with core dump, if enabled)
> 2: normal R exit
> 3: exit R without saving workspace
> 4: exit R saving workspace
> Selection: 3
>
> I conclude:
>  I am potentially running up against a limit in R, which should give a
> reasonable error, but currently just segfaults.
>
> My question:
> Most of the content of the JSON is an approximately 100K x 6K JSON
> equivalent of a dataframe, and I know R can handle much bigger than this
> size. I am expecting these JSON files to get even larger. My R code lives
> in a bigger system, and the JSON comes in via stdin, so I have absolutely
> no control over the data format. I can imagine trying to incrementally
> parse the JSON so I don't bump up against the limit, but I am eager for
> suggestions of simpler solutions.
>
> Also, I apologize for the timing of this bug report, as I know folks are
> working to get out the next release of R, but like so many things I have no
> control over when bugs leap up.
>
> Thanks.
>
> Jen
>
> > sessionInfo()
> R version 3.4.1 (2017-06-30)
> Platform: x86_64-pc-linux-gnu (64-bit)
> Running under: Ubuntu 14.04.5 LTS
>
> Matrix products: default
> BLAS: R-3.4.1/lib/libRblas.so
> LAPACK:R-3.4.1/lib/libRlapack.so
>
> locale:
>  [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C
>  [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8
>  [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8
>  [7] LC_PAPER=en_US.UTF-8       LC_NAME=C
>  [9] LC_ADDRESS=C               LC_TELEPHONE=C
> [11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C
>
> attached base packages:
> [1] stats     graphics  grDevices utils     datasets  methods   base
>
> loaded via a namespace (and not attached):
> [1] compiler_3.4.1
>
>         [[alternative HTML version deleted]]
>
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>
	[[alternative HTML version deleted]]

Thank you for your suggestion. Unfortunately, while R doesn't segfault
calling readr::read_file() on the test file I described, I get the error
message:

Error in read_file_(ds, locale) : negative length vectors are not allowed

Jen

On Sat, Sep 2, 2017 at 1:38 PM, Ista Zahn <istazahn at gmail.com> wrote:
> As s work-around I  suggest readr::read_file.
>
> --Ista
>
>
> On Sep 2, 2017 2:58 PM, "Jennifer Lyon" <jennifer.s.lyon at
gmail.com> wrote:
>
>> Hi:
>>
>> I have a 2.1GB JSON file. Typically I use readLines() and
>> jsonlite:fromJSON() to extract data from a JSON file.
>>
>> When I try and read in this file using readLines() R segfaults.
>>
>> I believe the two salient issues with this file are
>> 1). Its size
>> 2). It is a single line (no line breaks)
>>
>> I can reproduce this issue as follows
>> #Generate a big file with no line breaks
>> # In R
>> > writeLines(paste0(c(letters, 0:9), collapse=""),
"alpha.txt", sep="")
>>
>> # in unix shell
>> cp alpha.txt file.txt
>> for i in {1..26}; do cat file.txt file.txt > file2.txt && mv
-f file2.txt
>> file.txt; done
>>
>> This generates a 2.3GB file with no line breaks
>>
>> in R:
>> > moo <- readLines("file.txt")
>>
>>  *** caught segfault ***
>> address 0x7cffffff, cause 'memory not mapped'
>>
>> Traceback:
>>  1: readLines("file.txt")
>>
>> Possible actions:
>> 1: abort (with core dump, if enabled)
>> 2: normal R exit
>> 3: exit R without saving workspace
>> 4: exit R saving workspace
>> Selection: 3
>>
>> I conclude:
>>  I am potentially running up against a limit in R, which should give a
>> reasonable error, but currently just segfaults.
>>
>> My question:
>> Most of the content of the JSON is an approximately 100K x 6K JSON
>> equivalent of a dataframe, and I know R can handle much bigger than
this
>> size. I am expecting these JSON files to get even larger. My R code
lives
>> in a bigger system, and the JSON comes in via stdin, so I have
absolutely
>> no control over the data format. I can imagine trying to incrementally
>> parse the JSON so I don't bump up against the limit, but I am eager
for
>> suggestions of simpler solutions.
>>
>> Also, I apologize for the timing of this bug report, as I know folks
are
>> working to get out the next release of R, but like so many things I
have
>> no
>> control over when bugs leap up.
>>
>> Thanks.
>>
>> Jen
>>
>> > sessionInfo()
>> R version 3.4.1 (2017-06-30)
>> Platform: x86_64-pc-linux-gnu (64-bit)
>> Running under: Ubuntu 14.04.5 LTS
>>
>> Matrix products: default
>> BLAS: R-3.4.1/lib/libRblas.so
>> LAPACK:R-3.4.1/lib/libRlapack.so
>>
>> locale:
>>  [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C
>>  [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8
>>  [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8
>>  [7] LC_PAPER=en_US.UTF-8       LC_NAME=C
>>  [9] LC_ADDRESS=C               LC_TELEPHONE=C
>> [11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C
>>
>> attached base packages:
>> [1] stats     graphics  grDevices utils     datasets  methods   base
>>
>> loaded via a namespace (and not attached):
>> [1] compiler_3.4.1
>>
>>         [[alternative HTML version deleted]]
>>
>> ______________________________________________
>> R-devel at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-devel
>>
>
	[[alternative HTML version deleted]]

2017-09-02 20:58 GMT+02:00 Jennifer Lyon <jennifer.s.lyon at
gmail.com>:
> Hi:
>
> I have a 2.1GB JSON file. Typically I use readLines() and
> jsonlite:fromJSON() to extract data from a JSON file.
>
> When I try and read in this file using readLines() R segfaults.
>
> I believe the two salient issues with this file are
> 1). Its size
> 2). It is a single line (no line breaks)
As a workaround you can pipe something like "sed s/,/,\\n/g" before
your R script to insert line breaks.

I?aki

On Sat, Sep 2, 2017 at 8:58 PM, Jennifer Lyon <jennifer.s.lyon at
gmail.com> wrote:
> I have a 2.1GB JSON file. Typically I use readLines() and
> jsonlite:fromJSON() to extract data from a JSON file.
If your data consists of one json object per line, this is called
'ndjson'. There are several packages specialized to read ndjon files:

 - corpus::read_ndjson
 - ndjson::stream_in
 - jsonlite::stream_in

In particular the 'corpus' package handles large files really well
because it has an option to memory-map the file instead of reading all
of its data into memory.

If the data is too large to read, you can preprocess it using
https://stedolan.github.io/jq/ to extract the fields that you need.

You really don't need hadoop/spark/etc for this.

Jeroen:

Thank you for pointing me to ndjson, which I had not heard of and is
exactly my case.

My experience:
jsonlite::stream_in - segfaults
ndjson::stream_in - my fault, I am running Ubuntu 14.04 and it is too old
      so it won't compile the package
corpus::read_ndjson - works!!! Of course it does a different simplification
     than jsonlite::fromJSON, so I have to change some code, but it works
     beautifully at least in simple tests. The memory-map option may be of
     use in the future.

Another correspondent said that strings in R can only be 2^31-1 long, which
is why any "solution" that tries to load the whole file into R first
as a
string, will fail.

Thanks for suggesting a path forward for me!

Jen

On Sun, Sep 3, 2017 at 2:15 AM, Jeroen Ooms <jeroenooms at gmail.com>
wrote:
> On Sat, Sep 2, 2017 at 8:58 PM, Jennifer Lyon <jennifer.s.lyon at
gmail.com>
> wrote:
> > I have a 2.1GB JSON file. Typically I use readLines() and
> > jsonlite:fromJSON() to extract data from a JSON file.
>
> If your data consists of one json object per line, this is called
> 'ndjson'. There are several packages specialized to read ndjon
files:
>
>  - corpus::read_ndjson
>  - ndjson::stream_in
>  - jsonlite::stream_in
>
> In particular the 'corpus' package handles large files really well
> because it has an option to memory-map the file instead of reading all
> of its data into memory.
>
> If the data is too large to read, you can preprocess it using
> https://stedolan.github.io/jq/ to extract the fields that you need.
>
> You really don't need hadoop/spark/etc for this.
>
	[[alternative HTML version deleted]]