readJDX: Import a File Written in the JCAMP-DX Format

View source: R/readJDX.R

readJDXR Documentation

Import a File Written in the JCAMP-DX Format


This function supervises the entire import process. Not all official formats are supported, see the vignettes. Prior to release, this package is checked against a very large number of files in the author's collection. However, the JCAMP-DX standard allows many variations and it is difficult to anticipate all permutations. Error messages will generally let you know what's going on. If you have a file that you feel should be supported but gives an error, please file an issue at Github and share the file.


readJDX(file = "", SOFC = TRUE, debug = 0)



Character. The file name to import.


Logical. "Stop on Failed Check" The default is TRUE i.e. stop when something is not right. This ensures that correct data is returned. Change to FALSE at your own risk. NOTE: Only certain checks can be skipped via this option, as there are some parameters that must be available and correct in order to return any answer. For instance, one must end up with the same number of X and Y values. This option is provided for those advanced users who have carefully checked their original files and want to skip the required checks. It may also be useful for troubleshooting. The JCAMP-DX standard typically requires several checks of the data as it is decompressed. These checks are essential to obtaining the correct results. However, some JCAMP-DX writing programs do not follow the standard to the letter. For instance we have observed that not all JCAMP-DX writers put FIRSTY into the metadata, even though it is required by the standard. In other cases values in the file have low precision (see section on precision). Another example we have observed is NMR files where the X values are the count/index of data points, and FIRSTY is given in Hz. Since the field strength and center of the sweep frequency are needed to convert to ppm, and these are parameters not required in the standard, one cannot return an answer in either ppm or Hz automatically. In cases like this, one can set SOFC = FALSE and then manually convert the X axis.


Integer. The level of debug reporting desired. For those options giving a lot of output, you may wish to consider directing the output via sinkall and then search the results for the problematic lines.

  • 1 or higher = import progress is reported.

  • 2 = details about the variable lists, compression formats and parameters that were found.

  • 3 = print the extracted X values (huge output!).

  • 4 = detailed info on the Y value processing (huge output!).

  • 5 = detailed info about processing the Y values when DUP is in use (huge output!).

  • 6 = detailed info about processing the Y values when DIF is in use (huge output!).

In cases when an error is about to stop execution, you get additional information regardless of the debug value.


A list, as follows:

  • The first element is a data frame summarizing the pieces of the imported file.

  • The second element is the file metadata.

  • The third element is a integer vector giving the comment lines found (exclusive of the metdata, which typically contains many comments).

Additional elements contain the extracted data as follows:

  • If the file contains one non-NMR spectrum, or a processed NMR spectrum (i.e. only the final real data), a single data frame.

  • If the file contains the real and imaginary parts of a 1D NMR spectrum, there will be two data frames, one containing the real portion and the other the imaginary portion.

  • In all cases above, the data frame has elements x and y.

  • In the case of 2D NMR data, additional named list elements are returned including the F2 frequency values, the F1 frequency values, and a matrix containing the 2D data.

  • In the case of LC-MS or GC-MS data, a data frame is returned for each time point. The data frame has elements mz and int (intensity). Each time point is named with the time from the file.

Included Data Files

The examples make use of data files included with the package:

  • File SBO.jdx is an IR spectrum of Smart Balance Original spread (a butter substitute). The spectrum is presented in transmission format, and was recorded on a ThermoFisher instrument. The file uses AFFN compression, and was written with the JCAMP-DX 5.01 standard. Note that even though the Y-axis was recorded in percent transmission, in the JDX file it is stored on [0...1].

  • File PCRF.jdx is a 1H NMR spectrum of a hexane extract of a reduced fat potato chip. The spectrum was recorded on a JEOL instrument. The file uses SQZ DIF DUP compression, and was written with the JCAMP-DX 6.00 standard.

  • File PCRF_line265.jdx has a deliberate error in it.

  • File isasspc1.jdx is a 2D NMR file recorded on a JEOL GX 400 instrument. The file is freely available at

  • File MiniDIFDUP.JDX is a small demonstration file, used in the vignettes to illustrate the decompression process. It is derived from a freely available file.


Internally, this package uses a tolerance factor when comparing values during certain checks. This is desirable because the original values in the files are text strings of varying lengths which get converted to numerical values by R. Occasionally values in the file, such as FIRSTY, are stored with low precision, and the computation of the value to be compared occurs with much greater precision. In these cases the check can fail even when the tolerance is pretty loose. In these cases one might consider setting SOFC = FALSE to allow the calculation to proceed. If you do this, be certain to check the results carefully as described under SOFC.

Y Value Check

The standard requires a "Y Value Check" when in DIF mode. Extra Y values have been appended to each line to use in the check, and the last Y value on a line must equal the first Y value on the next line IFF one is in DIF mode. After a successful check, the extra Y value must be removed. In actual practice, some vendors, at least some of the time, seem to differ as to the meaning of "being in DIF mode". In turn, this determines how the Y value check should proceed.

  • The standard says "When, and only when, the last ordinate of a line is in DIF form ... The first ordinate of the next line ... is always an actual value, equal to the last calculated ordinate of the previous line". See section 5.8.3 of the 1988 publication.

  • Taking this definition literally, the Y value check (and removal of the extra value), should occur when one sees e.g. ... DIF DIF DIF (end of line). Let's call this "literal DIF". A literal DIF is easy to detect and act on.

  • In other cases, something like ... DIF DUP DUP (end of line) is considered to be in DIF mode for Y value check purposes. In these cases we have look backwards to see if we are in DIF mode. Let's call this "relayed DIF".

  • However, some vendors may treat ... DIF DUP DUP (end of line) as not in DIF mode, and hence one should not do the Y value check and not remove any values, as this vendor would not have added an extra Y value.

  • In addition to these three possibilities, readJDX through versions 0.3.xx used a different definition, namely if there were any DIF entries anywhere on the line, then DIF mode was assumed and the Y value check carried out. This worked for many files, but not all.

  • In the 0.4.xx series, readJDX detects both the literal and relayed definitions and tries to keep moving forward as much as possible.


readJDX is not particularly fast. Priority has been given to assuring correct answers, helpful debugging messages and understandable code.

See Also

Do browseVignettes("readJDX") for background information, references, supported formats, and details about the roles of each function. If you have a multiblock file (which contains multiple spectra, but not 2D NMR, LC-MS or GC-MS data sets), please see splitMultiblockDX for a function to break such files into individual files which can then be processed in the normal way.


# IR spectrum
sbo <- system.file("extdata", "SBO.jdx", package = "readJDX")
chk <- readJDX(sbo)
plot(chk[[4]]$x, chk[[4]]$y / 100,
  type = "l", main = "Original Smart Balance Spread",
  xlab = "wavenumber", ylab = "Percent Transmission"

# 1H NMR spectrum
pcrf <- system.file("extdata", "PCRF.jdx", package = "readJDX")
chk <- readJDX(pcrf)
plot(chk[[4]]$x, chk[[4]]$y,
  type = "l", main = "Reduced Fat Potato Chip Extract",
  xlab = "ppm", ylab = "Intensity"

# Capturing processing for troubleshooting
mdd <- system.file("extdata", "MiniDIFDUP.JDX", package = "readJDX")
tf <- tempfile(pattern = "Troubleshooting", fileext = "txt")
chk <- readJDX(mdd, debug = 6)
sinkall() # close the file connection

# 2D HETCORR spectrum
## Not run: 
nmr2d <- system.file("extdata", "isasspc1.dx", package = "readJDX")
chk <- readJDX(nmr2d)
contour(chk$Matrix, drawlabels = FALSE) # default contours not optimal

## End(Not run)

## Not run: 
# Line 265 has an N -> G error.  Try with various levels of debug.
# Even with debug = 0 you get useful diagnostic info.
problem <- system.file("extdata", "PCRF_line265.jdx", package = "readJDX")
chk <- readJDX(problem)

## End(Not run)

readJDX documentation built on Nov. 19, 2023, 1:08 a.m.