Error using textscan Badly formed format string??

Hello, I'm trying to plot the csv file (attached) but if keeps giving me the error:

"Error using textscan
Badly formed format string.
Error in johnvelasco_hw_3 (line 9).."

(line 9) which is..: 

dataArray = textscan(fileID, formatSpec, 'Delimiter', delimiter, 'EmptyValue', NaN,'HeaderLines', startRow-1, 'ReturnOnError', false);\

the complete code is

clear all;
close all;
clc;
filename = 'C:\Users\jvelasco2009\Downloads\hw3\CO-OPS__8722670__wl.csv';
delimiter = ',';
startRow = 2;
formatSpec = '%{yyyy-mm-dd HH:mm}D%f%f%f%f%f%f%g%{^\n\r}';
fileID = fopen(filename, 'r');
dataArray = textscan(fileID, formatSpec,  ...
                    'Delimiter', delimiter, ...
                    'EmptyValue', NaN, ...
                    'HeaderLines', startRow-1, ...
                    'ReturnOnError', false);
fclose(fileID);
DateTime = dataArray{:,1};
WaterLevel = dataArray{:, 2};
Sigma = dataArray {:, 3};
O = dataArray{:,4};
F = dataArray{:,5};
R = dataArray{:,6};
L = dataArray{:,7};
Quality = dataArray{:,8};
clearvars filename delimiter startRow formatSpec fileID dataArray one;
x = DateTime;
y = WaterLevel;
plot (x,y)
xlabel('Date and Time (08/01/2015-08/31/2015)')
ylabel('Water Level')
title('Tide Chart Lake Worth Pier, FL')
Mean = mean(WaterLevel)

I'd really appreciate the help, im [sic] stuck and I gotta turn this in soon. thank you!

Answer 1

[dpb]

formatSpec = '%{yyyy-mm-dd HH:mm}D%f%f%f%f%f%f%g%{^\n\r}';

The curly brackets don't belong around the trailing %{^\n\r} to skip remainder of line; use square brackets there instead. There are six numeric fields following the date column in the file which match the %f but the last column is character so the %g will fail. If you don't want that column, I'd use %*s for it altho the skip rest of line will work if just delete the %g.

To debug, these kinds of things, copy a line of the file to a string variable in the command window and then can use textscan on that string to debug a little at a time by parsing pieces of the string at a time.

formatSpec = '[%{yyyy-mm-dd HH:mm}D' repmat('%f',1,6) '%*s'];

Testing here (altho my release doesn't include the datetime class so I read the date fields as strings)...

>> s='08/31/2015 7:42  0.031  0.084  0  0  0  0  v';
>> fmt = ['%s %s' repmat('%f',1,6) '%*s'];
>> textscan(s,fmt,'collectoutput',1)
ans = 
  {1x2 cell}    [1x6 double]
>> ans{:}
ans = 
  '08/31/2015'    '7:42'
ans =
  0.0310    0.0840         0         0         0         0
>>

Voila!!! NB: I also used 'collectoutput' to minimize the cell arrays into two instead of several...

[Walter Roberson]

Also note that %D formats are new with R2014b so if you have an older release you will not be able to use %D formats.

[dpb]

That was why I read as strings--I presume the D format string descriptor was introduced with the datetime class, Walter, although it could theoretically have been applied for conversion directly to date numbers even as doubles--would been useful on occasion there as well.