(setf (vellum:at *table* 5 'first-column) "first value") Also, this means that you don't have to import symbols just to access data.įurthermore, VELLUM:AT is actually a setfable place which means that you can even. (print (vellum:at *table* 5 "first-column")) => :NULL oh, btw, column names are actually strings, meaning that this. (print (vellum:at *table* 5 'first-column)) => :NULL (print (vellum:row-count *table*)) => 0īut an attempt to get non-existing data won't signal error. This means that although, you can locate the last non-empty row. Data frames are sparse in concept, meaning that they lack a natural bound. However, the meaning of empty is somewhat blurred in vellum. (:name second-column :type fixnum)))) 0.5Read and write dataĪ freshly constructed data frame is obviously empty. (vellum:make-table :columns '(first-column You may be under impression that there has to be a better way to it. :header (vellum:make-header '(:name first-column) The header can be thought of as a data schema, containing column names, and column types for stored values and is required to create a table. 0.4Header + Columns = Data Frameĭata Frame is composed of header and columns. However, I hope to demonstrate that Vellum is a approachable library.
You are probably considering this is a drawback. Although I attempted to emulate R lang workflow, Common Lisp, even with all of its awesomeness, lacks some of the R features (namely first-class expressions and environments) therefore the end result can be considered to be quite a bit unique. The main goal of the Vellum was to create a lispy data frame library, and therefore it should come with no surprise that the design is very different from a more famous Pandas for python. Vellum is a such library for me, and I sincerely hope that it is a useful tool for you as well! 0.3Mindset I needed a nice, easy to use data frame library.
Define vellum code#
However, in combination, and especially in an interactive programming context when you want to get your code running as soon as possible, they really add up to a frustrating environment. None of the above issues alone is substantial enough to be considered a real problem. Enough so actually that I may simply run out of memory to represent all of my giganulls. Finally, I often work with a very sparse, and rather large, datasets. Higher-level semantics would help with that. After getting back to an old file I had to explain each individual line back to myself and this means that I am very likely to introduce mistakes If I have to change a thing. Secondly, I found that code written around such representations is difficult to read. First off, it is not very flexible: I don't want to keep track of what a column with a specific number holds, I just want to refer to it by name. This works but has a few distinct shortcomings. Then I would define accessors to obtain specific value from the data structure. Also, I simply like writing Lisp code.īefore Vellum existed I usually represented my data as vector-of-vectors (outer vector hold rows, inner vectors are rows). Although practice paints a less favorable picture, there are still use cases where Common Lisp could be considered to be a very good pick. This, in theory, makes is very much suited toward data processing and analysis. 0.2Why I made Vellum?Ĭommon Lisp is a fine language with exceptional interactivity, multiple implementations providing the potential for acceptable performance, useful features like condition system, and a stable standard base.
Define vellum software#
Note that this software is in beta quality at the moment. It is designed to be easy to both learn and use (especially in interactive fashion), practical, and reasonably efficient. Vellum is a data frame library build for Common Lisp.
0 kommentar(er)
