I have continuous sapflow data (averaged for every 10 minutes) in degree celsius recorded by heat-dissipation sensors for 30 trees (30 sapflow average value columns). But for proper analysis and calculation of sapflow data, TREX package is required. Since I haven't used it before, I had some error due to improper data structure. My data couldn't pass the "is.trex()" function. I have uploaded a small section of my dataset.

#data format
> timestamp            SF_Avg.1.   SF_Avg.2.   SF_Avg.3.   SF_Avg.4.   
1       TS               DegC       DegC          DegC      DegC      
2                         Avg        Avg           Avg       Avg       
3 2024-12-25 17:10:00   4.096447  9.202514      7.594289  5.357638  
4 2024-12-25 17:20:00   4.093174  9.271017      7.700016  5.424657  
5 2024-12-25 17:30:00   4.328177  9.294633      7.721289  5.534774   
6 2024-12-25 17:40:00   4.510964  9.441677      7.854154  5.740962 


#usage of code
>?is.trex
>>""is.trex(data, tz = 'UTC', tz.force = FALSE, time.format = '%m/%d/%y %H:%M:%S',
  solar.time = TRUE, long.deg = 7.7459, ref.add = FALSE, df = FALSE)""

How can I structure my data to pass this code?

I have continuous sapflow data (averaged for every 10 minutes) in degree celsius recorded by heat-dissipation sensors for 30 trees (30 sapflow average value columns). But for proper analysis and calculation of sapflow data, TREX package is required. Since I haven't used it before, I had some error due to improper data structure. My data couldn't pass the "is.trex()" function. I have uploaded a small section of my dataset.

#data format
> timestamp            SF_Avg.1.   SF_Avg.2.   SF_Avg.3.   SF_Avg.4.   
1       TS               DegC       DegC          DegC      DegC      
2                         Avg        Avg           Avg       Avg       
3 2024-12-25 17:10:00   4.096447  9.202514      7.594289  5.357638  
4 2024-12-25 17:20:00   4.093174  9.271017      7.700016  5.424657  
5 2024-12-25 17:30:00   4.328177  9.294633      7.721289  5.534774   
6 2024-12-25 17:40:00   4.510964  9.441677      7.854154  5.740962 


#usage of code
>?is.trex
>>""is.trex(data, tz = 'UTC', tz.force = FALSE, time.format = '%m/%d/%y %H:%M:%S',
  solar.time = TRUE, long.deg = 7.7459, ref.add = FALSE, df = FALSE)""

How can I structure my data to pass this code?

• r
• data-structures
• analysis

• Have you read the help page for is.trex? It gives very clear requirements for the "data" argument. Perhaps you can provide your data, or a sample of it, so that people here can help you pinpoint the problem. – Edward Commented Mar 7 at 7:35
• yes, and my data has the "Timestamp" column and the sapflow columns but I just wanted to know the acceptable data structure since my data showed errors with "unused arguments" – Sadi Commented Mar 7 at 7:38
• Then I suggest you copy and paste the error into your edited question, together with a sample of the data using dput as specified in minimal reproducible example and here – Edward Commented Mar 7 at 7:46
• From your data sample, it looks as though you have not imported your data properly. The first two rows contain text that should have been part of the column names. This means the numerical columns will be in character format rather than numeric. If your data frame is called data then you can try data <- data[-(1:2),] to remove these two rows, and then data[2:5] <- lapply(data[2:5], as.numeric) to convert the numeric columns to actual numbers. – Allan Cameron Commented Mar 7 at 8:49

2 Answers 2

The is.trex function allows two formats for the "data" argument. Both must be data frames:

1. Two columns where the first column is character and called "timestamp" and the second column is numeric and called "value"; or

2. Four columns containing the year of measurements (integer), day of year (integer), hour (character), and the measurement "value" (integer).

The help page states this very clearly. Your data does not meet either of these two formats for a few reasons. Firstly, there is no column called "value". Secondly, the columns are not in the required classes.

To fix the first problem, I would edit the source data by manually removing the second and third rows (or you can do that in R, but it's better to get the data types correct from the start). Assuming you've done this, you can then import the data, followed by a quick check:

sapflow <- read.table("sapflow.txt", header=TRUE)
head(sapflow)

            timestamp SF_Avg.1. SF_Avg.2. SF_Avg.3. SF_Avg.4.
1 2024-12-25 17:10:00  4.096447  9.202514  7.594289  5.357638
2 2024-12-25 17:20:00  4.093174  9.271017  7.700016  5.424657
3 2024-12-25 17:30:00  4.328177  9.294633  7.721289  5.534774
4 2024-12-25 17:40:00  4.510964  9.441677  7.854154  5.740962

The timestamp looks ok, but there is no "value" column. You can use the transform function to create a new column called "value" from each of the other columns one by one and then run the is.trex command separately.

SF1 <- transform(sapflow, value=SF_Avg.1.)
SF2 <- transform(sapflow, value=SF_Avg.2.) # etc.

is.trex(SF1, tz = 'UTC', ...) # etc.
2024-03-25 17:40:00 2024-03-25 17:50:00 2024-03-25 18:00:00 2024-03-25 18:10:00 
           4.096447            4.093174            4.328177            4.510964

Remove the header rows (TS, Avg) and keep only the timestamp and sapflow columns

data \<- read.csv("your_data.csv", skip = 2)
names(data) <- c("timestamp", paste0("SF", 1:30))

Ensure timestamp is in the correct format

data$timestamp <- format(as.POSIXct(data$timestamp), "%m/%d/%y %H:%M:%S")

Then try the is.trex function

library(TREX)
trex_data <- is.trex(data, 
                     tz = 'UTC',
                     time.format = '%m/%d/%y %H:%M:%S',
                     solar.time = TRUE,
                     long.deg = 7.7459)