Is there a way to filter, by first grep for a word in the content of the notes, and then with the result do a search on the content of these filtered notes ?

[barcell1]

I noticed that the grep or live_grep work different that i though, it locks indeed the first word that i search , retrieving the notes that contains it, but then the second type it search on the name/title of the notes, instead of searching in the content again.

Example what im trying to do:
i have like 4k notes, and suppose that i want to search a note that i know that contains the word "todo" and the word "job". i would like to find a way that i could search for "todo" and the retrivied notes that i got with that search, i would like to do another search for "job" in theirs content.

[ibhagwan]

Start with live grep, search for TODO, press ctrl-g to switch to grep and search for job this is the equivalent of :FzfLua grep search=TODO query=job.

[ibhagwan]

Rg/grep (the underlying command) is line based, you’re only getting the lines that match not full file content.

thanks, this explains everything !

Is there another way achieve this ? like a way in grep project to do this below : "dirpenal" (space) "menoridade" (i know this dont work, just wondering if there is a way) with fuzzy: dirpenal menoridade

dont work

You can put all the files of interest (via files or live_grep in the quick fix with alt-q) and then use grep_quickfix to search only these files.

[barcell1]

You can put all the files of interest (via files or live_grep in the quick fix with alt-q) and then use grep_quickfix to search only these files.

probably gonna work, but too much workaround, would be this a good feature request ? i mean would be feasible somehow having 2 full file content search concatenated for be implemented one day ?

[ibhagwan]

You can put all the files of interest (via files or live_grep in the quick fix with alt-q) and then use grep_quickfix to search only these files.

probably gonna work, but too much workaround, would be this a good feature request ? i mean would be feasible somehow having 2 full file content search concatenated for be implemented one day ?

There are many different ways to achieve this maybe with custom rg_opts, if the files are open you can also use lines picker - I don’t think another custom picker is required.

[barcell1]

There are many different ways to achieve this maybe with custom rg_opts, if the files are open you can also use lines picker - I don’t think another custom picker is required.

chatgpt create this function for me that looks like is working, you put the words separated by space.

local function grep_files_by_words()
  vim.ui.input({ prompt = "search files with words separated by space:  " }, function(input)
    if not input or input == "" then return end

    local words = vim.split(input, "%s+")
    if #words == 0 then return end

    local cwd = vim.fn.expand("~/PATH TO THE NOTES")
    local all_files = vim.fn.globpath(cwd, "**/*.md", false, true)

    local filtered = {}

    for _, file in ipairs(all_files) do
      local content = table.concat(vim.fn.readfile(file), "\n")
      local match = true
      for _, word in ipairs(words) do
        if not content:lower():find(word:lower(), 1, true) then
          match = false
          break
        end
      end
      if match then
        table.insert(filtered, file)
      end
    end

    if #filtered == 0 then
      vim.notify("No files found: " .. table.concat(words, ", "), vim.log.levels.INFO)
      return
    end

    require("fzf-lua").fzf_exec(filtered, {
      prompt = "Files with: " .. table.concat(words, " "),
      cwd = cwd,
      previewer = "bat_native",

      actions = {
        ["default"] = function(selected)
          if selected and selected[1] then
            vim.cmd("edit " .. vim.fn.fnameescape(selected[1]))
          end
        end,
      },
    })
  end)
end