The number of combinations of “text” or “ascii” data files that humans and computer programs can create is huge. Reading text files is far more complex than reading fixed format data bars from a data feed. Therefore, we cannot guarantee that every file you obtain or create will read properly into NeuroShell.

The help file has a topic called Data Files. This topic lays out the data formats that have been programmed, including the acceptable formats for dates and times. However, the following describes what types of bars can be read in those formats.

NeuroShell currently only recognizes tick, 1-second, 2-second, 5-second, 10-second, 20-second, 30-second, 1-minute, 2-minute, 5-minute, 10-minute, 15-minute, 20-minute, 30-minute, hourly and daily, weekly, monthly, and yearly text files. It will NOT recognize any other odd timed files like 7-second, 13-minute or 4-hour files.

However, if you load in a 1 minute file, you can specify a 7 or 13 minute chart, and the chart will be so created. The same is true if you load a 1 hour file and specify that you want a 3 or 6 hour chart.

Basically, then, 1-hour ASCII files can be graphed as 1-hour, 2-hour, 3-hour, X-hour charts as well as daily, weekly and monthly. 1-minute charts can be graphed as 1-minute, 2-minute, 3-minute, X-minute as well as X-hour, daily, weekly and monthly.

You can load tick files and format into anything you want (range, volume, second, minute, etc. charts). However, tick data files tend to be huge, and can slow up NeuroShell quite a bit.

The code that identifies a file’s frequency was very difficult to write, and has been modified as time has gone by to make it better. (We probably should have done it with a neural network, but we didn’t). If the file has errors, missing data, or other anomalies, it might not be recognized or read correctly. Sparse (irregular) and tick by tick data is especially difficult. We certainly cannot guarantee such data will load.

For speed purposes, NeuroShell currently only analyzes the first 50 rows of an ASCII file, so if the first 50 rows are in disarray (e.g., pre market data that is sparse and has no rhyme or reason to the time stamps), then it is entirely possible that NeuroShell could have trouble categorizing it. However, if the first 50 rows are nicely organized time-wise, NeuroShell will have a much better chance of correctly recognizing the time frequency.

NeuroShell won’t load range bar files specifically. However, if it happens to classify a range bar file as a tick file, then it MAY load, even though this was not specifically programmed. However if it analyzes the file and decides that it is some other frequency (for instance 1-minute), then you couldn’t make a range bar chart out of it.

To maximize ascii (text) file flexibility, we highly recommend that users create 1-day, 1-hour, 1-minute or tick charts and then let NeuroShell convert them to the proper frequency. So, for instance, instead of making a 5-minute file which can only be graphed in 5-minute increments, it would be better to make a 1-minute file and then have the flexibility to graph in any X-minute frequency. Additionally, if the user wants Volume, range, second bar charts, they are better off creating a tick by tick file (but not too big!) and letting NeuroShell convert the ticks to the proper frequency rather than trying to create a X volume, X range or X tick file and hoping that NeuroShell accidentally recognizes it (which NeuroShell isn’t programmed to do).