R interface to the peatland mid infrared database
pmird is an interface to the peatland mid infrared spectra database (pmird database) (Teickner, Pebesma, and Knorr 2025). ‘pmird’ allows to access the pmird database as dm object.
The pmird package can be installed as follows:
remotes::install_github("henningte/pmird")
The pmird database can be downloaded from Zotero (Teickner, Pebesma, and Knorr 2025). The downloaded database needs to be imported in a running MariaDB instance. In a linux terminal, the downloaded sql file can be imported like so:
Here,
The database itself does not contain the infrared spectra. These data are in folder pmird_prepared_data which needs to be stored at any place in the file system.
Once the database is set up and runs in a ‘MariaDB’ instance, it can be accessed from within R, using the ‘RMariaDB’ package (Müller et al. 2021):
library(pmird) library(RMariaDB) library(magrittr) library(ir) #> Registered S3 methods overwritten by 'tibble': #> method from #> format.tbl pillar #> print.tbl pillar library(quantities) #> Loading required package: units #> Warning: package 'units' was built under R version 4.0.5 #> udunits database from /usr/share/xml/udunits/udunits2.xml #> Loading required package: errors # connect to database con <- RMariaDB::dbConnect( drv = RMariaDB::MariaDB(), dbname = "pmird", default.file = "~/my.cnf", groups = "rs-dbi" )
Here, my.cnf is a text file that stores user and password information for the database server.
From this on, the ‘pmird’ R package can be used to access the database. The ‘pmird’ R package makes use of the R package ‘dm’ (Schieferdecker, Müller, and Bergant 2022) to access and manipulate the database contents. pmird::pm_get_dm() creates a dm object which stores the database structure.
# create the dm object dm_pmird <- pmird::pm_get_dm(con, learn_keys = TRUE)
The option learn_keys = TRUE means that information on the primary and foreign key are added to the dm object.
A dm object is a representation of the entire database and allows comfortable manipulation of the database from within R (e.g. addition of new rows to tables, addition of new tables, data queries) (Schieferdecker, Müller, and Bergant 2022).
Use case: obtaining general information on the datasets contained in the ‘pmird’ database
General information on the datasets contained in the ‘pmird’ database are stored in table datasets. The ‘pmird’ R package provides a function to obtain this table from the dm object (pmird::pm_get_table(.table_name = "datasets)).
# extract the datasets table pmird_datasets <- dm_pmird %>% pmird::pm_get_table(.table_name = "datasets")
This table can, for example, be used to select studies for which to extract data from the ‘pmird’ database.
Use case: extracting data for a specific dataset from the ‘pmird’ database
Assume you are interested in viewing all measured data for one specific dataset, e.g. the dataset with id_dataset == 8 in pmird_datasets. Using the ‘dm’ package and the dm object representing the ‘pmird’ database (dm_pmird), this data can be obtained as follows:
# get data for the dataset with ID 8 d8 <- dm_pmird %>% dm::dm_zoom_to(datasets) %>% dm::filter(id_dataset == 8) %>% dm::left_join(samples, by = "id_dataset") %>% dm::left_join(data_to_samples, by = "id_sample") %>% dm::left_join(data, by = "id_measurement") %>% dm::left_join(mir_metadata, by = "id_measurement") %>% dm::left_join(macrofossils, by = "id_measurement") %>% dm::pull_tbl() %>% tibble::as_tibble()
The resulting data frame (d8) contains information on all samples and measurements for the dataset with id_dataset == 8. d8 does not yet contain any spectra, but stores only information on the respective file paths to the downloaded spectra files within the downloaded folder pmird_prepared_data.
To load the spectra, you can use the function pmird::pmird_load_spectra(). In addition, you have to specify via argument directory in which folder pmird_prepared_data is stored (on the computer used for this tutorial, this was ".."):
# load the spectra d8 <- pmird::pm_load_spectra(d8, directory = "..")
pm_load_spectra() is a wrapper function around functions from the package ‘ir’ (Teickner 2022) which in turn are wrappers around read.spc() (Beleites and Sergo 2021) and read.csv(). pm_load_spectra() therefore can load spectra both saved as spc and csv files. pm_load_spectra() also converts d8 into an object of class ir from the ‘ir’ package. ‘ir’ provides functions for spectral preprocessing and manipulation (Teickner 2022) and is compatible with the ‘irpeat’ package which provides functions to analyze peat MIRS and spectral prediction models to predict peat properties from MIRS (Teickner and Hodgkins 2022).
Use case: Handling units and measurement errors
The R package ‘qauntities’ (Pebesma, Mailund, and Hiebert 2016; Ucar, Pebesma, and Azcorra 2018) can be used to add units and measurement errors to measured variables from the ‘pmird’ database. The ‘pmird’ R package allows batch unit and measurement error assignment:
# add information on units and errors with the 'quantities' package d8 <- d8 %>% pmird::pm_add_quantities() # show some values head(d8$N, 3) #> Units: [g/g] #> Errors: 0.0000996441 0.0001036096 0.0001168346 #> [1] 0.014054321 0.007934011 0.021408554
Use case: Generating data citations
Whoever uses data from the ‘pmird’ database must cite, in addition to the database, the original data sources for the used datasets. To make this straightforward, the ‘pmird’ R package contains the function pm_get_citations() to generate such a citation list for any extracted data subset:
# collect all citations for `d8`: d8_citations <- pm_get_citations( con = con, x = d8$id_measurement, file = "d8_citations.bib" ) #> Loading required namespace: bib2df # close connection to database RMariaDB::dbDisconnect(con)
The function takes column id_measurement of the extracted data and collects citations for all relevant data sources from the database. The results are exported to a bibtex file which is defined via argument file. This file can be imported to literature reference software. In this case, since the data have not been previously published, the created bibtex file is empty. RMariaDB::dbDisconnect(con) closes the connection to the database, as this is the final use case presented here.
Contributions
We welcome contributions from everyone. Please note that the pmird project is released with a Contributor Code of Conduct. By contributing to this project, you agree to abide by its terms.
References
Beleites, Claudia, and Valter Sergo. 2021. “hyperSpec: A Package to Handle Hyperspectral Data Sets in R.”
Müller, Kirill, Jeroen Ooms, David James, Saikat DebRoy, Hadley Wickham, and Jeffrey Horner. 2021. “RMariaDB: Database Interface and ’MariaDB’ Driver.”
Pebesma, Edzer, Thomas Mailund, and James Hiebert. 2016. “Measurement Units in R.” R Journal 8 (2): 486–94. https://doi.org/10.32614/RJ-2016-061.
Schieferdecker, Tobias, Kirill Müller, and Darko Bergant. 2022. “dm: Relational Data Models.”
Teickner, Henning. 2022. “ir: Functions to Handle and Preprocess Infrared Spectra.” Zenodo. https://doi.org/10.5281/ZENODO.6644806.
Teickner, Henning, and Suzanne Hodgkins. 2022. “irpeat 0.2.0: Functions to Analyze Mid-Infrared Spectra of Peat Samples.” Zenodo. https://doi.org/10.5281/ZENODO.7262744.
Teickner, Henning, Edzer Pebesma, and Klaus-Holger Knorr. 2025. “Peatland Mid Infrared Spectra Database.”
Ucar, Iñaki, Edzer Pebesma, and Arturo Azcorra. 2018. “Measurement Errors in R.” R Journal 10 (2): 549–57. https://doi.org/10.32614/RJ-2018-075.