Over the years of using JMP I have collected more than a few data tables. I tend to store the data tables in the same folder as the project that they are relevant to. I try to give the data table a meaningful name, such that I can remember later what the purpose of the data table was.
Over time, and even for the same project, the number of data tables with embedded graphs and scripts can start to grow. At some point it becomes harder top retrieve the data table that I used for a specific purpose.
When in contract, I write R scripts, I use copious amounts of commenting within the R script. And the script points to the data anyhow.
Does a mechanism exist in JMP that would allow me to embed (should be searchable) metadata about the project in the data table? For instance:
• project name
• customer name
• product name
• objective of the analysis
• description of data source
Within a JMP table, you could add a table script, that contained all of the comments that you normally add to your R scripts. Now the issue is that it would not be searchable using the Operating System types of searches.
However, what you could be doing is to use JMP Projects. These will allow you to create directory structures that contain all of the information that you are using for a given project. You could then easily adopt a process where you add a .txt file into each project directory, that details all of the issues that you comment about in your R code. And since JMP knows all about the files within a project, and the files remain as standard OS files to the Operating System, you have both worlds covered.
I add parameters to JMP datasets as table variables. Not sure how searchable this is. For example:
// Add new table variables showing the drug and relevant dates
snp_dataset << new table variable("Drug Name:", generic_drug_name);
snp_dataset << new table variable("Prev Start Date:", previous_start_date_txt);
snp_dataset << new table variable("Prev End Date:", previous_end_date_txt);
snp_dataset << new table variable("Curr Start Date:", current_start_date_txt);
snp_dataset << new table variable("Curr End Date:", current_end_date_txt);
The solution they provided is a JSL add-in that:
•Works for multiple file types.
•Is easy to maintain.
•Displays important metadata for any file through an interactive searchable interface.
•Lets the user add/change metadata and add files.
•Requires a minimal amount of user set-up.