Question

我的数据框如下所示：

 Name     Value   Value1  C1   C2    C3
  A        1       2      NA   NA    NA
  A        NA      2      NA   2     NA
  A        1       2      NA   2     NA
  A        1       2      NA   2     NA
  A        1       2      NA   2     NA
  B        NA      1      NA   2     NA
  B        NA      2      NA   2     NA
  B        1       NA     NA   2     NA
  B        1       NA     NA   2     NA
  C        1       5      NA   2     NA
  C        1       5      NA   2     NA

  dt <- as.data.table(df)
  new <- dt[is.na(`Value`) == FALSE & is.na(`Value1`) == FALSE,]

会给我以下输出：

 Name     Value   Value1  C1   C2    C3
  A        1       2      NA   NA    NA
  A        1       2      NA   2     NA
  A        1       2      NA   2     NA
  A        1       2      NA   2     NA
  C        1       5      NA   2     NA
  C        1       5      NA   2     NA

有没有办法做到这一点，但我希望我的列Name中的所有字母都不符合我的过滤条件，而不是过滤行。所以在这个例子中，我想回到A和B，因为A有一行没有两列Value和Value1非na值，B根本不符合标准。

Answer 1

您可以使用!作为否定运算符。（事实上，您应该已经使用它而不是== FALSE。此外，除非您的列名很奇怪（它们有空格或其他内容），否则您不需要反引号。

# your code, rewritten with ! instead of == FALSE
df[!is.na(Value) & !is.na(Value1), ]

# negate it to get the opposite
# same as above, but with !( your logic )
df[!(!is.na(Value) & !is.na(Value1)), ]
   Name Value Value1 C1 C2 C3
1:    A    NA      2 NA  2 NA
2:    B    NA      1 NA  2 NA
3:    B    NA      2 NA  2 NA
4:    B     1     NA NA  2 NA
5:    B     1     NA NA  2 NA

Answer 2

我可能误解了你，但就我可以阅读说明而言，我认为你的意思是获得特定Name的所有行，其中包含至少一个NA。

可以使用ave来完成此操作，我们会检查NA或Value列中是否至少有一个Value1，然后选择整个组Name }）。

df[with(df, ave(is.na(Value) | is.na(Value1), Name, FUN = any)), ]


 #  Name Value Value1 C1 C2 C3
 #1    A     1      2 NA NA NA
 #2    A    NA      2 NA  2 NA
 #3    A     1      2 NA  2 NA
 #4    A     1      2 NA  2 NA
 #5    A     1      2 NA  2 NA
 #6    B    NA      1 NA  2 NA
 #7    B    NA      2 NA  2 NA
 #8    B     1     NA NA  2 NA
 #9    B     1     NA NA  2 NA

Answer 3

@RonakShah已经指出（使用public class GridWidgetService extends RemoteViewsService { @Override public RemoteViewsFactory onGetViewFactory(Intent intent) { return new NewsListRemoteVieFactory(this.getApplicationContext()); } } class NewsListRemoteVieFactory implements RemoteViewsService.RemoteViewsFactory{ public static ArrayList<News> newsArrayList = new ArrayList<>(); Context mContext; public NewsListRemoteVieFactory(Context applicationContext) { this.mContext = applicationContext; } private void readNews(){ SharedPreferences prefs = PreferenceManager.getDefaultSharedPreferences(mContext.getApplicationContext()); Gson gson = new Gson(); String json = prefs.getString("news", ""); Type type = new TypeToken<ArrayList<News>>(){}.getType(); newsArrayList = gson.fromJson(json, type); } @Override public void onCreate() { readNews(); } @Override public void onDataSetChanged() { readNews(); } @Override public void onDestroy() { } @Override public int getCount() { if(newsArrayList == null){ return 0; } return newsArrayList.size(); } @Override public RemoteViews getViewAt(int position) { RemoteViews views = new RemoteViews(mContext.getPackageName(), R.layout.widget_grid_view_item); views.setTextViewText(R.id.widget_grid_view_item, "\u2022 " + newsArrayList.get(position).getTitle() + "\n" + String.valueOf(newsArrayList.get(position).getDescription())); Bundle selectedNewsBundle = new Bundle(); selectedNewsBundle.putParcelableArrayList("news",newsArrayList); Intent fillInIntent = new Intent(); views.setOnClickFillInIntent(R.id.widget_grid_view_item, fillInIntent); //PendingIntent pendingIntent = PendingIntent.getBroadcast(mContext, 0, fillInIntent, PendingIntent.FLAG_UPDATE_CURRENT); //views.setOnClickPendingIntent(R.id.widget_grid_view_item, pendingIntent); Intent intent = new Intent(mContext, MainActivity.class); PendingIntent pendingIntent = PendingIntent.getActivity(mContext, 0, intent, 0); views.setOnClickPendingIntent(R.id.widget_grid_view_item, pendingIntent); return views; } @Override public RemoteViews getLoadingView() { return null; } @Override public int getViewTypeCount() { return 1; } @Override public long getItemId(int position) { return position; } @Override public boolean hasStableIds() { return true; } }）如何base-R filter Name行NA或Value中至少有Value1行{1}}列。

但是，或许，OP正在寻找仅unique个名称而不是完整的行子集。在dplyr中，可以这样做：

library(dplyr)

df %>% group_by(Name) %>%
  filter_at(vars(starts_with("Value")), any_vars(is.na(.))) %>%  #Either Value or Value1 
  select(Name) %>% distinct()

# Name 
# <chr>
# 1 A    
# 2 B

要获取Name或NA或Value

中Value1

df %>% group_by(Name) %>%
  filter(any(is.na(Value) | any(is.na(Value1))))

  #   Name  Value Value1 C1       C2 C3   
  #  <chr> <int>  <int> <lgl> <int> <lgl>
  # 1 A         1      2 NA       NA NA   
  # 2 A        NA      2 NA        2 NA   
  # 3 A         1      2 NA        2 NA   
  # 4 A         1      2 NA        2 NA   
  # 5 A         1      2 NA        2 NA   
  # 6 B        NA      1 NA        2 NA   
  # 7 B        NA      2 NA        2 NA   
  # 8 B         1     NA NA        2 NA   
  # 9 B         1     NA NA        2 NA

的完整行

df <- read.table(text = 
"Name     Value   Value1  C1   C2    C3
A        1       2      NA   NA    NA
A        NA      2      NA   2     NA
A        1       2      NA   2     NA
A        1       2      NA   2     NA
A        1       2      NA   2     NA
B        NA      1      NA   2     NA
B        NA      2      NA   2     NA
B        1       NA     NA   2     NA
B        1       NA     NA   2     NA
C        1       5      NA   2     NA
C        1       5      NA   2     NA",
header = TRUE, stringsAsFactors = FALSE)

数据：

public

如果其他两列不符合过滤条件，则从列中获取值

3 个答案: