Downloading data from txt file

Question

Jesus Christian 2017-10-24

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/363036-downloading-data-from-txt-file

评论： Cedric 2017-10-26

I was trying to download data from txt file using the importdata function however it downloads it into a structure with cell arrays in it. this is the code I am using. I need this data in a matrix of doubles if possible but it won't work.

delimiterIn = ' ';
[DataStructure] = importdata('iss.txt',delimiterIn);
DataCells = DataStructure(1).textdata(:,:);

This is a sample of what the txt file has inside:

25544U 98067A   16001.21631944  .00005382  00000-0  85171-4 0  9992
25544  51.6436 175.7680 0008248 349.1433 350.2600 15.55191287978846
25544U 98067A   16001.28471065  .00005383  00000-0  85187-4 0  9993
25544  51.6436 175.4264 0008246 349.3894  13.1693 15.55192096978860
25544U 98067A   16001.49479696  .00005567  00000-0  87829-4 0  9993
25544  51.6434 174.3772 0008262 350.0277 109.5268 15.55195766978890
25544U 98067A   16001.49479696  .00005567  00000-0  87829-4 0  9993
25544  51.6434 174.3772 0008262 350.0277 109.5268 15.55195766978890
.
.
.

2 个评论
显示无隐藏无

Cedric 2017-10-24

编辑：Cedric 2017-10-24

Which data do you need? What about the ones with letters/dashes? Do you want to discard them, to extract them in a separate array, ..? They have 9 columns actually, whereas the others have 8.

Jesus Christian 2017-10-25

I need the entire bottom row of every set in one array of 8 elements (i.e. the rows with a 2 as first element)

I also need the top row (rows with 1 as a first element) but only the elements without letters

Basically, I need all the elements on each row that don't have letters.

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Cedric 2017-10-25

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/363036-downloading-data-from-txt-file#answer_287568

编辑：Cedric 2017-10-25

在 MATLAB Online 中打开

data.txt

Here is one way:

 data = textscan( fileread( 'data.txt' ), '%s' ) ;
 data = reshape( str2double( data{1} ), 17, [] ).' ;
 data = data(:,~isnan( data(1,:))) ;

Then it is possible to split into two blocks if you prefer, for rows "1" and rows "2".

PS: it would be possible to extract non-numeric data as well if you needed.

6 个评论
显示 4更早的评论隐藏 4更早的评论

Cedric 2017-10-25

编辑：Cedric 2017-10-25

在 MATLAB Online 中打开

There is a discrepancy after line 1500. I copied lines 1499 to 1502 here:

25544U 98067A   16136.69611507  .00005252  00000-0  85076-4 0  9996
25544  51.6433 220.1152 0002135 107.8767 350.8535 15.54602232999921
25544U 98067A   16137.21209581  .00005359  00000-0  86643-4 0  9999
25544  51.6435 217.5406 0002111 109.7961 358.5795 15.54608822    13

You can see that lines 1499 and 1500 follow the 9-8 columns pattern present from the beginning of the file, but that lines 1501 and 1502 are 9-9. In fact, from line 1502 on, you loose regularity. The plot below shows the # of elements per line:

You can see that it is regular (dense) 8-9 until 1502. It seems that there are other, minor discrepancies in your data set. Sometimes there is a + in front of positive numbers, sometimes not. Sometimes there is a minus (do you really have negative values?). Also, looking at values between the 8 and 9 values versions of lines with ID 2, it seems that this is a filling problem:

25544  51.6433 220.9943 0002124 108.0375  83.7509 15.54599868999892
25544  51.6433 220.9943 0002124 108.0375  83.7509 15.54599868999892
25544  51.6433 220.1152 0002135 107.8767 350.8535 15.54602232999921
25544  51.6435 217.5406 0002111 109.7961 358.5795 15.54608822    13
25544  51.6440 215.8562 0001995 108.3968  92.4187 15.54614828    61
25544  51.6440 215.8562 0001995 108.3968  92.4187 15.54614828    61

You'll have to understand what is going wrong data-wise, and then we can find a way to process/import the content.

Here is how I looked up for the discrepancy by the way:

 >> rows = strsplit( strtrim( fileread( 'iss.txt' )), '\n' ) ;
 >> rows = cellfun( @strsplit, rows, 'Unif', false ) ;
 >> rowSz = cellfun( 'length', rows ) ;
 >> plot( rowSz ) ;
 >> pos = strfind( rowSz, [9,9] ) ; % Look for first break of [9,8] pattern.
 >> pos(1)
 ans =
        1501

Walter Roberson 2017-10-26

Looks to me as if possibly some of the fields are fixed width -- that perhaps none of the lines are only 8 fields and instead the fields are running together. If so, then that second last field can be handled with %.12f

Cedric 2017-10-26

You have at least to understand the structure of the data that you are reading, and tell us if, as Walter suggests, there are 9 columns everywhere with fixed width or if it is something else. That must be documented somewhere, and especially the length of the last field.

请先登录，再进行评论。

Downloading data from txt file

2 个评论
显示无隐藏无

回答（1 个）

6 个评论
显示 4更早的评论隐藏 4更早的评论

另请参阅

类别

标签

Community Treasure Hunt

Downloading data from txt file

2 个评论 显示 无隐藏 无

回答（1 个）

6 个评论 显示 4更早的评论隐藏 4更早的评论

另请参阅

类别

标签

Community Treasure Hunt

2 个评论
显示无隐藏无

6 个评论
显示 4更早的评论隐藏 4更早的评论