Reading csv file NaN and NaT problem

SKG
2018 10 31
3 个回答
更新时间：2018 10 31
29 次查看（30 天）

0 个投票

Hi all,
I am trying to make a csv file read.File is located in "file path". The Variable Names is in line 2.
detectImportOptions(filepath);
opts.DataLines=[1,Inf];
T = readtable(filepath,opts);
I can read the file but the problem is that some of the variables are changing to NaT or NaN, is it possible to set on line 2 (where the variable names are located) to character before executing readable function?

请先登录，再回答此问题。

回答（3 个）

Jeremy Hughes
Jeremy Hughes 2018-10-31

0 个投票

Based on the description, I think this is what you'd want. Without an attached file, I cannot say more.
opts = detectImportOptions(filepath);
opts.VariableNamesLine = 2;
opts.DataLines=[3,Inf];
T = readtable(filepath,opts,'ReadVariableNames',true);
dpb
dpb 2018-10-31

0 个投票

detectImportOptions(filepath);
You didn't save the output of detectImportOptions so you don't have an import options object at this point.
Hence,
opts.DataLines=[1,Inf];
simply creates the one element in the struct named opts that you passed to readtable but everything else is its default, NOT what the detectImportOptions function would have divined to be the characteristics of the file.
Try
opts=detectImportOptions(filepath);
T = readtable(filepath,opts);
first before resorting to anything more exotic; generally the automagic scanning routines are able to infer "who's who in the zoo" without human intervention.
You can use setvartype, yes, if it turns out to be needed as well as setting the specific time format string, but try the above first.

5 个评论
显示 3更早的评论 隐藏 3更早的评论

SKG
SKG 2018-10-31
Actually I made a mistake when I was writing the question. I already tried that. It is not working. It is not taking the second line's all variables as characters.
dpb
dpb 2018-10-31
Well, show us a small section of the input file...
Jeremy Hughes
Jeremy Hughes 2018-10-31
There's no way to have one row be text and the others numeric. Tables are homogeneous in the "row" direction. This is why the fields are being converted to NaN or NaT.
Try my answer above and see if that gives you closer to what you expect.
SKG
SKG 2018-10-31
I already tried, it is not working
dpb
dpb 2018-10-31
编辑：dpb 2018-10-31
ONE MORE TIME!!!!
We can't help you if you won't show us what the input data format actually is.
Attach a few lines of the data file that illustrates the format and where you have a problem.

请先登录，再进行评论。

SKG
SKG 2018-10-31

0 个投票

Solved it:
opts = detectImportOptions(filepath); opts.DataLines=[1,Inf ]; opts.VariableTypes(1, 1:end) = {'char'};

1 个评论
显示 -1更早的评论 隐藏 -1更早的评论

Jeremy Hughes
Jeremy Hughes 2018-10-31
Note, this will set all the data you import to be cells of character vectors, no numbers.

请先登录，再进行评论。

类别

在 帮助中心 和 File Exchange 中查找有关 Data Import and Analysis 的更多信息

标签

提问:

SKG
SKG
2018-10-31

评论：

Jeremy Hughes
Jeremy Hughes
2018-10-31

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by